diff --git a/built-in-nodes/AnimaLLLiteApply.mdx b/built-in-nodes/AnimaLLLiteApply.mdx
index ac0853f7b..40ad940e4 100644
--- a/built-in-nodes/AnimaLLLiteApply.mdx
+++ b/built-in-nodes/AnimaLLLiteApply.mdx
@@ -14,13 +14,13 @@ AnimaLLLiteApply applies a lightweight animation patch to a diffusion model, ena
 |-----------|-------------|-----------|----------|-------|
 | `model` | The base diffusion model to apply the patch to | MODEL | Yes | |
 | `model_patch` | The pre-configured animation patch to apply | MODEL_PATCH | Yes | |
-| `image` | The reference image to guide the generation | IMAGE | Yes | |
+| `image` | The reference image to guide the generation. Only the first 3 color channels (RGB) are used | IMAGE | Yes | |
 | `strength` | The strength of the patch effect (default: 1.0) | FLOAT | Yes | -10.0 to 10.0 |
 | `start_percent` | The percentage of the denoising process at which the patch starts taking effect (default: 0.0) | FLOAT | Yes | 0.0 to 1.0 |
 | `end_percent` | The percentage of the denoising process at which the patch stops taking effect (default: 1.0) | FLOAT | Yes | 0.0 to 1.0 |
 | `mask` | An optional mask to limit the patch effect to specific areas of the image | MASK | No | |
 
-**Note on parameter constraints:** If the `model_patch` has 4 input channels and no `mask` is provided, a zero mask is automatically created to match the image dimensions. If the `model_patch` does not have 4 input channels, the `mask` parameter is ignored and set to `None`.
+**Note on parameter constraints:** If the `model_patch` has 4 input channels and no `mask` is provided, a zero mask is automatically created to match the image dimensions. If the `model_patch` does not have 4 input channels, the `mask` parameter is ignored and set to `None`. This node is marked as experimental in ComfyUI.
 
 ## Outputs
 
diff --git a/built-in-nodes/BerniniConditioning.mdx b/built-in-nodes/BerniniConditioning.mdx
index ff8b37759..5b2514944 100644
--- a/built-in-nodes/BerniniConditioning.mdx
+++ b/built-in-nodes/BerniniConditioning.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-The BerniniConditioning node prepares video and image conditioning data for the Wan2.2-A14B model. It encodes source videos, reference videos, and reference images using the provided VAE, then attaches them to the conditioning data for in-context generation tasks. The task is automatically inferred from which inputs are connected.
+The BerniniConditioning node prepares video and image conditioning data for the Wan2.2-A14B model. It encodes source videos, reference videos, and reference images using the provided VAE, then attaches them to the conditioning data as in-context tokens. The task is automatically inferred from which inputs are connected.
 
 ## Inputs
 
@@ -20,8 +20,8 @@ The BerniniConditioning node prepares video and image conditioning data for the
 | `length` | Number of frames in the output latent (default: 81) | INT | Yes | 1 to 8192 (step: 4) |
 | `batch_size` | Number of videos to generate in a single batch (default: 1) | INT | Yes | 1 to 4096 |
 | `source_video` | Source video to edit or restyle (v2v, rv2v). Resized to width/height and trimmed to length. | IMAGE | No | - |
-| `reference_video` | Video to insert into the source video (ads2v). | IMAGE | No | - |
-| `reference_images` | Reference images injected as in-context tokens (r2v, rv2v). Up to 8 images can be provided. | IMAGE | No | 0 to 8 images |
+| `reference_video` | Video to insert into the source video (ads2v). Trimmed to length and resized with preserved aspect ratio (long edge capped at ref_max_size). | IMAGE | No | - |
+| `reference_images` | Reference images injected as in-context tokens (r2v, rv2v). Each image is encoded independently at its own native aspect ratio (long edge capped at ref_max_size). Up to 8 images can be provided. | IMAGE | No | 0 to 8 images |
 | `ref_max_size` | Max size for the long edge of reference_video and reference_images. Resized with preserved aspect ratio and snapped to 16px (default: 848). | INT | No | 16 to 8192 (step: 16) |
 
 **Note:** The task is inferred from which inputs are connected:
@@ -31,6 +31,8 @@ The BerniniConditioning node prepares video and image conditioning data for the
 - `reference_images` only → reference-to-video (r2v)
 - `source_video` + `reference_video` → insert image/video into video (ads2v)
 
+Context streams are attached in a fixed order: `source_video` (source ID 1), then `reference_video` (source ID 2), then each reference image (source IDs 3, 4, ...). A reference image input containing multiple images contributes one stream per image.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/BuildJsonPromptIdeogram.mdx b/built-in-nodes/BuildJsonPromptIdeogram.mdx
index 6dca9115c..32be68390 100644
--- a/built-in-nodes/BuildJsonPromptIdeogram.mdx
+++ b/built-in-nodes/BuildJsonPromptIdeogram.mdx
@@ -13,14 +13,14 @@ This node constructs a structured JSON prompt specifically formatted for the Ide
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `element` | Prompt elements from the node Create Bounding Boxes. | ARRAY | Yes | - |
-| `high_level_description` | Optional description of the image in one or two sentences. Strongly recommended for better results. (default: empty) | STRING | No | - |
-| `background` | Description of the image background or environment. Providing a value is strongly recommended – the node will run with an empty string but results may be unpredictable. (default: empty) | STRING | No | - |
+| `high_level_description` | Optional description of the image in one or two sentences. Strongly recommended. (default: empty) | STRING | No | - |
+| `background` | Mandatory description of the image background or environment. Providing a value is strongly recommended – the node will run with an empty string but results may be unpredictable. (default: empty) | STRING | No | - |
 | `style` | The visual style category for the generated image. Select one of the options. (default: "none") | COMBO | Yes | `"none"`<br />`"photo"`<br />`"art_style"` |
 | `photo` | Camera or lens details for photographic outputs (e.g. 35mm, f/1.4, bokeh). Only available when `style` is set to "photo". (default: empty) | STRING | No | - |
 | `art_style` | Art style description (e.g. flat vector illustration, bold outlines). Only available when `style` is set to "art_style". (default: empty) | STRING | No | - |
-| `aesthetics` | Aesthetic keywords (e.g. moody, cinematic, desaturated). Providing a value is strongly recommended. (default: empty) | STRING | No | - |
-| `lighting` | Lighting description (e.g. golden hour, rim light, dramatic shadows). Providing a value is strongly recommended. (default: empty) | STRING | No | - |
-| `medium` | Medium type (e.g. photograph, illustration, 3d_render, painting, graphic_design). Providing a value is strongly recommended. (default: empty) | STRING | No | - |
+| `aesthetics` | Mandatory aesthetic keywords (e.g. moody, cinematic, desaturated). Providing a value is strongly recommended. (default: empty) | STRING | No | - |
+| `lighting` | Mandatory lighting description (e.g. golden hour, rim light, dramatic shadows). Providing a value is strongly recommended. (default: empty) | STRING | No | - |
+| `medium` | Mandatory medium type (e.g. photograph, illustration, 3d_render, painting, graphic_design). When style = photo, set to photograph. Providing a value is strongly recommended. (default: empty) | STRING | No | - |
 | `color_palette` | Hex color codes that steer the image's dominant colors. Up to 16 entries. | COLORS | No | - |
 
 **Note:** When `style` is set to "photo", the `photo` input becomes available and you should usually set `medium` to "photograph". When `style` is set to "art_style", the `art_style` input becomes available. The `background`, `aesthetics`, `lighting`, and `medium` parameters have default empty strings but are considered mandatory for reliable image generation; providing values is strongly recommended.
diff --git a/built-in-nodes/ByteDanceSeedAudio.mdx b/built-in-nodes/ByteDanceSeedAudio.mdx
index 133e08d34..af520641d 100644
--- a/built-in-nodes/ByteDanceSeedAudio.mdx
+++ b/built-in-nodes/ByteDanceSeedAudio.mdx
@@ -12,33 +12,33 @@ Generate speech, music, sound effects and multi-speaker dialogue from a single p
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `text_prompt` | Describe the voice(s), emotion, pacing, ambience, background music and sound effects, and include the lines to speak (name characters inline for dialogue). In 'audio reference' mode, refer to connected clips by order as @Audio1, @Audio2, @Audio3. Maximum 3000 characters. | STRING | Yes | 1 to 3000 characters |
-| `reference_mode` | How to condition the voice: 'text only' (describe everything in the prompt), 'audio reference' (clone up to 3 voices, tagged @Audio1-3), 'image reference' (derive a voice from one character image), or 'preset voice' (pick a built-in named voice that reads the prompt). | COMBO | Yes | `"text only"`<br />`"audio reference"`<br />`"image reference"`<br />`"preset voice"` |
-| `reference_audio_1` | Reference clip for voice cloning, tagged @Audio1 in the prompt. Up to 30s. Only available when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
-| `reference_audio_2` | Reference clip tagged @Audio2 in the prompt. Up to 30s. Only available when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
-| `reference_audio_3` | Reference clip tagged @Audio3 in the prompt. Up to 30s. Only available when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
-| `reference_image` | A single character image; the model derives a voice from it. Cannot be combined with reference audio. Only available when `reference_mode` is "image reference". | IMAGE | No | - |
-| `preset_voice` | A built-in TTS 2.0 voice that reads the prompt. No reference clip needed, and @AudioN tags are not used in this mode. Only available when `reference_mode` is "preset voice". | COMBO | No | Multiple options available (see description) |
+| `text_prompt` | Describe the voice(s), emotion, pacing, ambience, background music and sound effects, and include the lines to speak (name characters inline for dialogue). In "audio reference" mode, refer to connected clips by order as @Audio1, @Audio2, @Audio3. With the multilingual model, a quoted line can start with a timestamp range that controls when and how long it is spoken, e.g. `[5.5s:8.0s] Wait for me!`. Write the prompt in the same language as the lines to speak. Minimum 1 character, maximum 3000 characters. | STRING | Yes | 1 to 3000 characters |
+| `reference_mode` | How to condition the voice: "text only" (describe everything in the prompt), "audio reference" (clone up to 3 voices, tagged @Audio1-3), "image reference" (derive a voice from one character image), or "preset voice" (pick a built-in named voice that reads the prompt). | COMBO | Yes | `"text only"`<br />`"audio reference"`<br />`"image reference"`<br />`"preset voice"` |
+| `reference_audio_1` | Reference clip for voice cloning, tagged @Audio1 in the prompt. Up to 30s. Available only when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
+| `reference_audio_2` | Reference clip tagged @Audio2 in the prompt. Up to 30s. Available only when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
+| `reference_audio_3` | Reference clip tagged @Audio3 in the prompt. Up to 30s. Available only when `reference_mode` is "audio reference". | AUDIO | No | Up to 30 seconds |
+| `reference_image` | A single character image; the model derives a voice from it. Cannot be combined with reference audio. Available only when `reference_mode` is "image reference". | IMAGE | No | - |
+| `preset_voice` | A built-in TTS 2.0 voice that reads the prompt. No reference clip needed, and @AudioN tags are not used in this mode. Required when `reference_mode` is "preset voice". | COMBO | No | Multiple built-in preset voice options (first option selected by default) |
 | `sample_rate` | Output sample rate in Hz. (default: "24000") | COMBO | Yes | `"8000"`<br />`"16000"`<br />`"24000"`<br />`"32000"`<br />`"44100"`<br />`"48000"` |
 | `speech_rate` | Speaking speed. 0 = normal, 100 = 2.0x, -50 = 0.5x. (default: 0) | INT | Yes | -50 to 100 |
 | `loudness_rate` | Loudness. 0 = normal, 100 = 2.0x, -50 = 0.5x. (default: 0) | INT | Yes | -50 to 100 |
 | `pitch_rate` | Pitch shift in semitones (-12 to 12). (default: 0) | INT | Yes | -12 to 12 |
 | `seed` | Seed controls whether the node should re-run; results are non-deterministic regardless of seed. (default: 42) | INT | Yes | 0 to 2147483647 |
-| `model` | Model version. The multilingual variant supports 20 languages and per-sentence timing via `[5.5s:8.0s]` timestamps. (default: "seed-audio-1.0-multilingual") | COMBO | No | `"seed-audio-1.0-multilingual"`<br />`"seed-audio-1.0"` |
+| `model` | Model version. `seed-audio-1.0-multilingual` supports 20 languages and per-sentence timing control via `[5.5s:8.0s]` timestamps. `seed-audio-1.0` supports English and Chinese only, without timing control. (default: "seed-audio-1.0-multilingual") | COMBO | No | `"seed-audio-1.0-multilingual"`<br />`"seed-audio-1.0"` |
 
 ### Parameter Constraints
 
 - **Reference mode dependencies**: The `reference_mode` parameter determines which other inputs are required:
   - **"text only"**: No additional inputs required. The prompt must not contain @AudioN tags.
-  - **"audio reference"**: Requires at least one of `reference_audio_1`, `reference_audio_2`, or `reference_audio_3` to be connected. Reference clips must be connected in order without gaps (e.g., _1, then _2, then _3). Each clip is limited to 30 seconds maximum duration. The prompt must reference connected clips using @Audio1, @Audio2, @Audio3 tags.
-  - **"image reference"**: Requires `reference_image` to be connected. The prompt must not contain @AudioN tags.
-  - **"preset voice"**: Requires `preset_voice` to be selected. The prompt must not contain @AudioN tags (the entire prompt is read in the selected voice).
+  - **"audio reference"**: Requires at least one of `reference_audio_1`, `reference_audio_2`, or `reference_audio_3` to be connected. Reference clips must be connected in order without gaps. Each clip is limited to 30 seconds maximum duration. If @AudioN tags are used in the prompt, the highest tag number must not exceed the number of connected reference clips.
+  - **"image reference"**: Requires `reference_image` to be connected. @AudioN tags are not used; the prompt should contain only the text to synthesize.
+  - **"preset voice"**: Requires a preset voice to be selected. The whole prompt is read in the selected voice; @AudioN tags are not used as references, and tags such as @Audio2 or higher are rejected.
 
-- **Audio reference ordering**: When using "audio reference" mode, reference audio inputs must be connected sequentially starting from `reference_audio_1` without gaps. For example, you can connect _1 and _2, but not _1 and _3 without _2.
+- **Audio reference ordering**: In "audio reference" mode, reference audio inputs must be connected sequentially starting from `reference_audio_1` without gaps. For example, you can connect `reference_audio_1` and `reference_audio_2`, but not `reference_audio_1` and `reference_audio_3` without `reference_audio_2`.
 
-- **Maximum audio tags**: The prompt can reference up to 3 audio clips (@Audio1, @Audio2, @Audio3) when in "audio reference" mode. The highest numbered tag must not exceed the number of connected reference audio inputs.
+- **Maximum audio tags**: In "audio reference" mode, up to 3 reference clips can be connected (@Audio1, @Audio2, @Audio3), and the highest @AudioN tag in the prompt cannot exceed the number of connected reference audio inputs.
 
-- **Model differences**: The "seed-audio-1.0-multilingual" model supports 20 languages (English, Chinese, Japanese, Korean, Mexican & Castilian Spanish, Indonesian, German, Brazilian Portuguese, French, Thai, Vietnamese, Malay, Filipino, Italian, Russian, Dutch, Polish, Turkish, Swedish) plus per-sentence timing control using timestamps in the format `[5.5s:8.0s]`. The "seed-audio-1.0" model supports English and Chinese only, without timing control.
+- **Model differences**: The `seed-audio-1.0-multilingual` model supports 20 languages (English, Chinese, Japanese, Korean, Mexican & Castilian Spanish, Indonesian, German, Brazilian Portuguese, French, Thai, Vietnamese, Malay, Filipino, Italian, Russian, Dutch, Polish, Turkish, Swedish) plus per-sentence timing control using timestamps in the format `[5.5s:8.0s]`. The `seed-audio-1.0` model supports English and Chinese only, without timing control.
 
 ## Outputs
 
diff --git a/built-in-nodes/ConvertDictionaryToString.mdx b/built-in-nodes/ConvertDictionaryToString.mdx
index 5ab0a57f9..ffcc8cc22 100644
--- a/built-in-nodes/ConvertDictionaryToString.mdx
+++ b/built-in-nodes/ConvertDictionaryToString.mdx
@@ -6,8 +6,6 @@ icon: "circle"
 mode: wide
 ---
 
-Convert Dictionary to String
-
 This node converts a dictionary (a collection of key-value pairs) into a text string in JSON format. You can control the indentation level to make the output more readable or compact.
 
 ## Inputs
diff --git a/built-in-nodes/DA3GeometryToMesh.mdx b/built-in-nodes/DA3GeometryToMesh.mdx
index 4a916d972..52609d72f 100644
--- a/built-in-nodes/DA3GeometryToMesh.mdx
+++ b/built-in-nodes/DA3GeometryToMesh.mdx
@@ -13,13 +13,15 @@ This node converts a DA3_GEOMETRY packet into a 3D mesh by unprojecting the dept
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `da3_geometry` | The DA3_GEOMETRY packet containing depth map, optional confidence map, optional sky map, and source image | DA3_GEOMETRY | Yes | - |
-| `batch_index` | Which image of a batch to convert. Per-image vertex counts differ so batches cannot be stacked (default: 0) | INT | Yes | 0 to 4096 |
+| `batch_index` | Which image of a batch to convert. Per-image vertex counts differ so batches cannot be stacked (default: 0). Must be less than the batch size of the input geometry, otherwise an error is raised | INT | Yes | 0 to 4096 |
 | `decimation` | Vertex stride. 1 = full resolution, 2 = half, etc. (default: 1) | INT | Yes | 1 to 8 |
 | `discontinuity_threshold` | Drop triangles whose 3x3 depth span exceeds this fraction. 0 = off (default: 0.04) | FLOAT | Yes | 0.0 to 1.0 |
 | `confidence_threshold` | Exclude pixels whose per-image normalised confidence is below this value. 0 = keep all, 1 = keep only the single most confident pixel. Used when the geometry has a confidence map (Small/Base models) (default: 0.1) | FLOAT | Yes | 0.0 to 1.0 |
 | `use_sky_mask` | Exclude sky-probability pixels (sky >= 0.5) from the mesh. Used when the geometry has a sky map (Mono/Metric models) (default: True) | BOOLEAN | Yes | True or False |
 | `texture` | Use the source image as a base color texture (default: True) | BOOLEAN | Yes | True or False |
 
+Pixels with non-finite, zero, or negative depth values are always excluded from the mesh. An error is raised if the resulting mesh is empty; the error message suggests raising `discontinuity_threshold`, lowering `confidence_threshold`, or disabling `use_sky_mask`.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/DA3GeometryToPointCloud.mdx b/built-in-nodes/DA3GeometryToPointCloud.mdx
index 0b9e3c452..356d5fd65 100644
--- a/built-in-nodes/DA3GeometryToPointCloud.mdx
+++ b/built-in-nodes/DA3GeometryToPointCloud.mdx
@@ -18,11 +18,18 @@ Convert a depth map from a DA3_GEOMETRY object into a 3D point cloud. It applies
 | `use_sky_mask` | Exclude sky-probability pixels (sky >= 0.5). Used when the geometry has a sky map (Mono/Metric models). (default: True) | BOOLEAN | Yes | True or False |
 | `downsample` | Take every Nth pixel (1 = full resolution). Higher values give fewer points and faster processing. (default: 1) | INT | Yes | 1 to 16 |
 
+**Notes:**
+- `batch_index` must be smaller than the batch size of the input geometry; otherwise the node raises an error.
+- If filtering removes all pixels, the node raises an error. Lowering `confidence_threshold` or disabling `use_sky_mask` can prevent this.
+- `confidence_threshold` is only applied when the geometry includes a confidence map (Small/Base models).
+- `use_sky_mask` is only applied when the geometry includes a sky map (Mono/Metric models).
+- When `downsample` is greater than 1, the camera intrinsics are scaled accordingly so the projected points remain correct.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
-| `point_cloud` | A point cloud object containing filtered 3D points, optional colors, and optional confidence values | DA3_POINT_CLOUD |
+| `point_cloud` | A point cloud containing filtered 3D points, optional colors, and optional confidence values | DA3_POINT_CLOUD |
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/DA3GeometryToPointCloud/en.md)
 
diff --git a/built-in-nodes/DA3Inference.mdx b/built-in-nodes/DA3Inference.mdx
index d3aa32ae3..b3e87c7fd 100644
--- a/built-in-nodes/DA3Inference.mdx
+++ b/built-in-nodes/DA3Inference.mdx
@@ -21,10 +21,10 @@ This node runs the Depth Anything 3 model on an image to estimate depth and geom
 | `pose_method` | How the camera field-of-view is estimated (for Small and Base models only). cam_dec: learned from image features. ray_pose: derived geometrically from the model's 3D ray output. Affects perspective correctness of the 3D output (default: cam_dec) | COMBO | No (conditional) | `"cam_dec"`<br />`"ray_pose"` |
 
 **Notes on parameter constraints:**
-- The `ref_view_strategy` and `pose_method` parameters are only available when `mode` is set to `"multiview"`
-- Multi-view mode requires a Small or Base model variant. Models with other head types (such as Metric or Mono) do not support cross-view attention or camera pose estimation
-- When `pose_method` is set to `"cam_dec"`, the model must have a camera decoder. If set to `"ray_pose"`, the model must have a DualDPT head
-- If the selected `pose_method` is not compatible with the loaded model, an error will be raised
+- The `ref_view_strategy` and `pose_method` parameters are only available when `mode` is set to `"multiview"`.
+- Multi-view mode requires a Small or Base model variant. Models with other head types (such as Metric or Mono) do not support cross-view attention or camera pose estimation.
+- When `pose_method` is set to `"cam_dec"`, the model must have a camera decoder. If set to `"ray_pose"`, the model must have a DualDPT head.
+- If the selected `pose_method` is not compatible with the loaded model, an error will be raised.
 
 ## Outputs
 
diff --git a/built-in-nodes/DA3Render.mdx b/built-in-nodes/DA3Render.mdx
index f0413e2de..64630761a 100644
--- a/built-in-nodes/DA3Render.mdx
+++ b/built-in-nodes/DA3Render.mdx
@@ -13,7 +13,7 @@ This node renders a visualization from Depth Anything 3 geometry data. It can ou
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `da3_geometry` | The Depth Anything 3 geometry data packet containing depth, and optionally sky and confidence tensors | DA3_GEOMETRY | Yes | - |
-| `output` | The type of visualization to render. Options include depth, depth_colored, sky_mask, and confidence. Each option has its own set of sub-parameters. | COMBO | Yes | `"depth"`<br />`"depth_colored"`<br />`"sky_mask"`<br />`"confidence"` |
+| `output` | The type of visualization to render.<br />- depth: normalised greyscale depth image.<br />- depth_colored: depth mapped through the Turbo colormap.<br />- sky_mask: sky probability in [0, 1] (for Mono/Metric models only).<br />- confidence: normalised depth confidence (for Small/Base models only). | COMBO | Yes | "depth"<br />"depth_colored"<br />"sky_mask"<br />"confidence" |
 
 ### Sub-parameters for `output` options
 
@@ -21,8 +21,8 @@ When `output` is set to `"depth"` or `"depth_colored"`:
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `normalization` | The depth normalization method. v2_style uses mean/std normalization for perceptually balanced results (default). min_max stretches the full depth range to [0, 1] for maximum contrast. raw preserves metric units for Metric model without scaling. | COMBO | Yes | `"v2_style"`<br />`"min_max"`<br />`"raw"` |
-| `apply_sky_clip` | Clip sky-region depth to the 99th percentile of foreground depth before normalization. Requires a sky key in the da3_geometry input (for Mono/Metric models only). Default: False | BOOLEAN | Yes | True<br />False |
+| `normalization` | Depth normalization method.<br />- v2_style: mean/std normalisation for perceptually balanced results (default).<br />- min_max: stretches the full depth range to [0, 1] for maximum contrast.<br />- raw: no scaling, preserves metric units for Metric model. | COMBO | Yes | "v2_style"<br />"min_max"<br />"raw" |
+| `apply_sky_clip` | Clip sky-region depth to the 99th percentile of foreground depth before normalisation. Requires a sky key in the da3_geometry input (for Mono/Metric models only). Default: False | BOOLEAN | Yes | True<br />False |
 
 When `output` is set to `"sky_mask"`:
 
diff --git a/built-in-nodes/GeminiVideoOmni.mdx b/built-in-nodes/GeminiVideoOmni.mdx
index 65ae23a81..72bd1d115 100644
--- a/built-in-nodes/GeminiVideoOmni.mdx
+++ b/built-in-nodes/GeminiVideoOmni.mdx
@@ -20,6 +20,10 @@ Generate a video with audio from a text prompt using Google's Gemini Omni Flash
 | `temperature` | Controls randomness in generation (default: 1.0). | FLOAT | No | 0.0 to 2.0 |
 | `top_p` | Nucleus sampling parameter (default: 0.95). | FLOAT | No | 0.0 to 1.0 |
 
+Notes:
+- If an image input contains multiple frames, each frame counts toward the maximum of 14 images.
+- When `images` or `videos` are provided, the combined encoded media size must stay under about 90 MB; otherwise the node raises an error.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
@@ -30,4 +34,4 @@ Generate a video with audio from a text prompt using Google's Gemini Omni Flash
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/GeminiVideoOmni/en.md)
 
 ---
-**Source fingerprint (SHA-256):** `420f62e09e7744554f5989cac7754703fd6fd268a000b45b32d7487beac21c42`
+**Source fingerprint (SHA-256):** `1b7ca51d07cfb6a166cfed2a7e7174fd62f3290abcc1bdfdce94369dda242d3f`
diff --git a/built-in-nodes/HeyGenAvatarVideoNode.mdx b/built-in-nodes/HeyGenAvatarVideoNode.mdx
index a21090a44..d3aa41cad 100644
--- a/built-in-nodes/HeyGenAvatarVideoNode.mdx
+++ b/built-in-nodes/HeyGenAvatarVideoNode.mdx
@@ -13,28 +13,28 @@ Generate a talking-presenter video from a HeyGen avatar. This node creates a vid
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `engine` | Rendering engine; each choice lists only the avatars that support it. 'auto' offers every avatar and picks its best engine (Avatar IV preferred). Avatar V is highest fidelity, Avatar III is the most affordable. | COMBO | Yes | `"auto"`<br />`"avatar_iv"`<br />`"avatar_iii"`<br />`"avatar_v"` |
-| `avatar` | Avatar look to present the video. The available options depend on the selected engine – when engine is `"auto"` a curated list of HeyGen's public avatars is shown; for specific engines only avatars that support that engine are listed. | COMBO | Yes | Varies based on `engine` selection; see tooltip on the widget. |
+| `avatar` | Avatar look to present the video. With engine `"auto"`, it is curated from HeyGen's public library and the best engine the look supports is chosen automatically. With engine `"avatar_iv"`, `"avatar_iii"`, or `"avatar_v"`, only avatar looks that support that engine are listed. | COMBO | Yes | Varies based on the `engine` selection |
 | `custom_avatar_id` | Optional HeyGen avatar look ID. When set, overrides the avatar selected above. Any of HeyGen's 3000+ public looks (or your private avatars) can be used. | STRING | No |  |
 | `speech` | Drive the avatar with a text script (HeyGen text-to-speech) or your own audio. When `"script"` is chosen, see the sub-parameters below; when `"audio"` is chosen, provide an audio file for lip-sync. | COMBO | Yes | `"script"`<br />`"audio"` |
 | `resolution` | Output video resolution (default: `"1080p"`). | COMBO | No | `"720p"`<br />`"1080p"` |
 | `aspect_ratio` | Output aspect ratio. `"auto"` follows the avatar's source footage (default: `"auto"`). | COMBO | No | `"auto"`<br />`"16:9"`<br />`"9:16"`<br />`"1:1"`<br />`"4:5"`<br />`"5:4"` |
-| `background_color` | Optional solid background color as a hex code (e.g. `'#00ff00'`). Leave empty for the avatar's own background. If provided, must start with `#`. | STRING | No |  |
+| `background_color` | Optional solid background color as a hex code (e.g. `'#00ff00'`). Leave empty for the avatar's own background. If provided, the value must start with `#`. | STRING | No |  |
 | `seed` | Not sent to HeyGen; change it to force a re-run (default: `42`). | INT | No | Min: 0<br />Max: 2147483647 |
 
-### Sub‑parameters when `speech` is `"script"`
+### Sub-parameters when `speech` is `"script"`
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `text` | Text for the avatar to speak (up to 5000 characters). The generated speech must be at least 1 second long. | STRING | Yes |  |
-| `voice` | Voice for the script. The default option uses the voice HeyGen assigned to the avatar. | COMBO | Yes | Multiple options available (includes default voice plus HeyGen's general voice library). |
+| `voice` | Voice for the script. The default option uses the voice HeyGen assigned to the avatar. | COMBO | Yes | Multiple options available (default voice plus HeyGen's general voice library) |
 | `custom_voice_id` | Optional HeyGen voice ID. When set, overrides the voice selected above. Any voice from HeyGen's library (2000+) can be used. | STRING | No |  |
 | `voice_speed` | Speech speed multiplier (default: `1.0`). | FLOAT | No | Min: 0.5<br />Max: 1.5<br />Step: 0.05 |
 
-### Sub‑parameters when `speech` is `"audio"`
+### Sub-parameters when `speech` is `"audio"`
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `audio` | Audio for the avatar to lip‑sync, up to 10 minutes. | AUDIO | Yes |  |
+| `audio` | Audio for the avatar to lip-sync, up to 10 minutes. | AUDIO | Yes |  |
 
 ## Outputs
 
diff --git a/built-in-nodes/HeyGenCreateAvatarNode.mdx b/built-in-nodes/HeyGenCreateAvatarNode.mdx
index d341202cc..7ea2857da 100644
--- a/built-in-nodes/HeyGenCreateAvatarNode.mdx
+++ b/built-in-nodes/HeyGenCreateAvatarNode.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-Create a reusable HeyGen avatar from a photo of a person or from a text prompt that describes a character to generate. The resulting avatar ID can be used with the HeyGen Avatar Video node to create videos featuring this avatar.
+Create a reusable HeyGen avatar from a photo of a person or from a text prompt that describes a character to generate. The resulting avatar ID can be used with the HeyGen Avatar Video node to create videos featuring this avatar. Save the ID to reuse the avatar in future workflows.
 
 ## Inputs
 
@@ -18,8 +18,8 @@ When `source` is set to `"prompt"`, the following additional parameters become a
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `prompt` | Description of the avatar to generate (up to 1000 characters). | STRING | Yes | 1 to 1000 characters |
-| `reference_images` | Up to 3 reference images guiding the generated look. | IMAGE | No | 0 to 3 images |
+| `prompt` | Description of the avatar to generate (up to 1000 characters). Must contain at least 1 non-whitespace character. Default: empty string. | STRING | Yes | 1 to 1000 characters |
+| `reference_images` | Up to 3 reference images guiding the generated look. Downscaled automatically if larger than 2K. | IMAGE | No | 0 to 3 images |
 
 When `source` is set to `"photo"`, the following additional parameter becomes available:
 
diff --git a/built-in-nodes/HeyGenTalkingPhotoNode.mdx b/built-in-nodes/HeyGenTalkingPhotoNode.mdx
index f791aaf82..61aabdbf8 100644
--- a/built-in-nodes/HeyGenTalkingPhotoNode.mdx
+++ b/built-in-nodes/HeyGenTalkingPhotoNode.mdx
@@ -14,16 +14,18 @@ Animate a still image of a person into a lip-synced talking video using HeyGen's
 |-----------|-------------|-----------|----------|-------|
 | `image` | Image of a person to animate. Downscaled automatically if larger than 2K. | IMAGE | Yes | - |
 | `speech` | Drive the avatar with a text script (HeyGen text-to-speech) or your own audio. | COMBO | Yes | `"script"`<br />`"audio"` |
-| `text` | Text for the avatar to speak (up to 5000 characters). The generated speech must be at least 1 second long. | STRING | Yes (when speech is "script") | - |
-| `voice` | Voice for the script (HeyGen's most popular voices). | COMBO | Yes (when speech is "script") | Multiple options available |
-| `custom_voice_id` | Optional HeyGen voice ID. When set, overrides the voice selected above. Any voice from HeyGen's library (2000+) can be used. | STRING | No | - |
+| `text` | Text for the avatar to speak (up to 5000 characters). The generated speech must be at least 1 second long. (default: empty) | STRING | Yes (when speech is `"script"`) | - |
+| `voice` | Voice for the script (HeyGen's most popular voices). | COMBO | Yes (when speech is `"script"`) | Multiple options available |
+| `custom_voice_id` | Optional HeyGen voice ID. When set, overrides the voice selected above. Any voice from HeyGen's library (2000+) can be used. (default: empty) | STRING | No | - |
 | `voice_speed` | Speech speed multiplier (default: 1.0). | FLOAT | No | 0.5 to 1.5 |
-| `audio` | Audio for the avatar to lip-sync, up to 10 minutes. | AUDIO | Yes (when speech is "audio") | - |
-| `resolution` | Output video resolution (default: "1080p"). | COMBO | No | `"720p"`<br />`"1080p"` |
-| `aspect_ratio` | Output aspect ratio. 'auto' follows the input image (default: "auto"). | COMBO | No | `"auto"`<br />`"16:9"`<br />`"9:16"`<br />`"1:1"`<br />`"4:5"`<br />`"5:4"` |
-| `expressiveness` | How expressive the animated face and gestures are (default: "low"). | COMBO | No | `"low"`<br />`"medium"`<br />`"high"` |
+| `audio` | Audio for the avatar to lip-sync, up to 10 minutes. | AUDIO | Yes (when speech is `"audio"`) | - |
+| `resolution` | Output video resolution (default: `"1080p"`). | COMBO | No | `"720p"`<br />`"1080p"` |
+| `aspect_ratio` | Output aspect ratio. `"auto"` follows the input image (default: `"auto"`). | COMBO | No | `"auto"`<br />`"16:9"`<br />`"9:16"`<br />`"1:1"`<br />`"4:5"`<br />`"5:4"` |
+| `expressiveness` | How expressive the animated face and gestures are (default: `"low"`). | COMBO | No | `"low"`<br />`"medium"`<br />`"high"` |
 | `seed` | Not sent to HeyGen; change it to force a re-run (default: 42). | INT | No | 0 to 2147483647 |
 
+Note: `speech` is a dynamic input that selects the speech source. When set to `"script"`, `text` and `voice` are required, and `custom_voice_id` and `voice_speed` are optional. When set to `"audio"`, `audio` is required instead.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/HeyGenTextToSpeechNode.mdx b/built-in-nodes/HeyGenTextToSpeechNode.mdx
index e8be62ce5..a0c9edfd7 100644
--- a/built-in-nodes/HeyGenTextToSpeechNode.mdx
+++ b/built-in-nodes/HeyGenTextToSpeechNode.mdx
@@ -14,7 +14,7 @@ Synthesize speech audio from text using HeyGen's Starfish TTS engine. This node
 |-----------|-------------|-----------|----------|-------|
 | `text` | Text to synthesize (up to 5000 characters). The generated speech must be at least 1 second long. | STRING | Yes | 1 to 5000 characters |
 | `voice` | Voice to use (curated from HeyGen's most popular Starfish-compatible voices). | STRING | Yes | Multiple options available |
-| `custom_voice_id` | Optional HeyGen voice ID. When set, overrides the voice selected above. The voice must support the Starfish engine. | STRING | No | Any valid HeyGen voice ID |
+| `custom_voice_id` | Optional HeyGen voice ID. When set, overrides the voice selected above. The voice must support the Starfish engine (default: empty). | STRING | No | Any valid HeyGen voice ID |
 | `speed` | Speech speed multiplier (default: 1.0). | FLOAT | No | 0.5 to 2.0 (step: 0.05) |
 | `ssml` | Treat the text as SSML markup (for pauses, emphasis, and pronunciation control) (default: False). | BOOLEAN | No | True / False |
 | `seed` | Not sent to HeyGen; change it to force a re-run (default: 42). | INT | No | 0 to 2147483647 |
diff --git a/built-in-nodes/HeyGenVideoTranslateNode.mdx b/built-in-nodes/HeyGenVideoTranslateNode.mdx
index 2fc791ab1..e9f77d6bc 100644
--- a/built-in-nodes/HeyGenVideoTranslateNode.mdx
+++ b/built-in-nodes/HeyGenVideoTranslateNode.mdx
@@ -14,9 +14,9 @@ Translate a spoken video into another language with voice cloning and lip sync.
 |-----------|-------------|-----------|----------|-------|
 | `video` | Video with speech to translate. | VIDEO | Yes | - |
 | `output_language` | Target language for the translated video. | STRING | Yes | "Arabic"<br />"Bengali"<br />"Chinese"<br />"Danish"<br />"Dutch"<br />"English"<br />"French"<br />"German"<br />"Greek"<br />"Hindi"<br />"Indonesian"<br />"Italian"<br />"Japanese"<br />"Korean"<br />"Malay"<br />"Polish"<br />"Portuguese"<br />"Russian"<br />"Spanish"<br />"Swedish"<br />"Tamil"<br />"Telugu"<br />"Thai"<br />"Turkish"<br />"Ukrainian"<br />"Vietnamese" |
-| `mode` | 'speed' is faster; 'precision' produces higher-quality lip sync at twice the price. | STRING | Yes | "speed"<br />"precision" |
+| `mode` | 'speed' is faster; 'precision' produces higher-quality lip sync at twice the price. (default: "speed") | STRING | Yes | "speed"<br />"precision" |
 | `translate_audio_only` | Only swap the audio track, keeping the original mouth movements (no lip sync). (default: False) | BOOLEAN | No | True<br />False |
-| `speaker_count` | Number of speakers in the video. 0 = detect automatically. (default: 0) | INT | No | 0 to 10 |
+| `speaker_count` | Number of speakers in the video. 0 = detect automatically. Values above 0 are sent to the API as the number of speakers. (default: 0) | INT | No | 0 to 10 |
 | `seed` | Not sent to HeyGen; change it to force a re-run. (default: 42) | INT | No | 0 to 2147483647 |
 
 ## Outputs
diff --git a/built-in-nodes/IdeogramPImage.mdx b/built-in-nodes/IdeogramPImage.mdx
new file mode 100644
index 000000000..efe421b3d
--- /dev/null
+++ b/built-in-nodes/IdeogramPImage.mdx
@@ -0,0 +1,34 @@
+---
+title: "IdeogramPImage - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the IdeogramPImage node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "IdeogramPImage"
+icon: "circle"
+mode: wide
+---
+
+Ideogram P-Image generates images from a text prompt using Ideogram's fast text-to-image model, which is known for strong typography and photorealism. It also supports Ideogram 4.0 structured JSON captions for exact control over text strings, colors, and layout. The node returns the generated image(s) along with the final prompt the image was actually generated from.
+
+## Inputs
+
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `prompt` | Text prompt. Also accepts an Ideogram 4.0 structured JSON caption (exact colors as #RRGGBB hexes, exact text strings, bounding-box layout) — set prompt_upsampling to OFF to use it verbatim. Must not be empty. (default: "") | STRING | Yes | Any text |
+| `quality` | Speed/price/quality tier. MEDIUM is the everyday default; HIGH for complex prompts, fine detail and difficult text; VERY_LOW/LOW for drafts at scale. Difficult text renders poorly below MEDIUM. (default: "MEDIUM") | STRING | Yes | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
+| `resolution` | Output size class (exact pixels follow the aspect ratio, e.g. 16:9 gives 1280x720 at 1K and 2560x1440 at 2K). Prefer HIGH + 2K for crisp typography. (default: "1K") | STRING | Yes | "1K"<br />"2K" |
+| `aspect_ratio` | The aspect ratio for image generation. (default: "1:1") | STRING | Yes | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
+| `prompt_upsampling` | Expands short prompts into a detailed structured caption before generation (the rewritten prompt is returned as final_prompt). Set OFF when supplying your own JSON caption or exact wording. (default: "AUTO") | STRING | Yes | "AUTO"<br />"ON"<br />"OFF" |
+| `seed` | Seed for reproducible generation. With prompt_upsampling OFF, the same seed and settings return the same image; with ON/AUTO the prompt rewrite varies per run — reproduce a result by reusing its final_prompt output with prompt_upsampling OFF and the same seed. (default: 42) | INT | No | 0 to 2147483647 |
+
+**Note on constraints:** The prompt must contain at least one non-whitespace character, otherwise the node fails. Set `prompt_upsampling` to OFF when supplying your own structured JSON caption or exact wording. When `prompt_upsampling` is ON or AUTO, the prompt is rewritten before generation, so the same seed may not reproduce the same image; to reproduce an image, reuse its `final_prompt` output with `prompt_upsampling` OFF and the same seed.
+
+## Outputs
+
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `image` | The generated image(s) returned as a batch of images. If Ideogram's content safety filter blocks the generation, an error is raised instead. | IMAGE |
+| `final_prompt` | The prompt the image was actually generated from (the rewritten structured caption when prompt_upsampling ran, else your prompt). Feed it back with prompt_upsampling OFF and the same seed to reproduce this image. | STRING |
+
+> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/IdeogramPImage/en.md)
+
+---
+**Source fingerprint (SHA-256):** `7bd20aae508fee111ded32e87119ed6fc01c5ad5ba7d595e24391830a0f20bb7`
diff --git a/built-in-nodes/KlingDualCharacterVideoEffectNode.mdx b/built-in-nodes/KlingDualCharacterVideoEffectNode.mdx
index 5b9066af0..34be9207e 100644
--- a/built-in-nodes/KlingDualCharacterVideoEffectNode.mdx
+++ b/built-in-nodes/KlingDualCharacterVideoEffectNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: Legacy model options (`kling-v1`, `kling-v1-5`, `kling-v1-6`) are scheduled for retirement on **September 15, 2026**. Use Kling 3.0 effect workflows instead. See [Kling 3.0 legacy retirement](/tutorials/partner-nodes/kling/kling-3-0#legacy-kling-models-eol).
-</Warning>
-
-The Kling Dual Character Video Effect Node creates videos with special effects based on the selected scene.
+The Kling Dual Character Video Effect Node creates videos with special effects based on the selected scene. It takes two images and positions the first image on the left side and the second image on the right side of the composite video. Different visual effects are applied depending on the chosen effect scene.
 
 ## Inputs
 
diff --git a/built-in-nodes/KlingImage2VideoNode.mdx b/built-in-nodes/KlingImage2VideoNode.mdx
index 83da4ced6..7af2e292c 100644
--- a/built-in-nodes/KlingImage2VideoNode.mdx
+++ b/built-in-nodes/KlingImage2VideoNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: Legacy model options (`kling-v1-5`, `kling-v1-6`, `kling-v2-1`, `kling-v2-master`, and related variants) are scheduled for retirement on **September 15, 2026**. Use Kling 3.0 or 2.6 models instead. See [Kling 3.0 legacy retirement](/tutorials/partner-nodes/kling/kling-3-0#legacy-kling-models-eol).
-</Warning>
-
-The Kling Image to Video Node generates a video from a starting reference image using text prompts.
+The Kling Image to Video Node generates a video from a starting reference image using text prompts. It takes an image as the first frame and creates a video sequence based on positive and negative text descriptions, with configurable options for model, duration, aspect ratio, and generation mode.
 
 ## Inputs
 
diff --git a/built-in-nodes/KlingSingleImageVideoEffectNode.mdx b/built-in-nodes/KlingSingleImageVideoEffectNode.mdx
index d7072a97d..f7d228f14 100644
--- a/built-in-nodes/KlingSingleImageVideoEffectNode.mdx
+++ b/built-in-nodes/KlingSingleImageVideoEffectNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: Legacy model options (`kling-v1-5`, `kling-v1-6`) are scheduled for retirement on **September 15, 2026**. Use Kling 3.0 effect workflows instead. See [Kling 3.0 legacy retirement](/tutorials/partner-nodes/kling/kling-3-0#legacy-kling-models-eol).
-</Warning>
-
-The Kling Single Image Video Effect Node creates videos with different special effects based on a single reference image.
+The Kling Single Image Video Effect Node creates videos with different special effects based on a single reference image. It applies various visual effects and scenes to transform static images into dynamic video content. The node supports different effect scenes, model options, and video durations to achieve the desired visual outcome.
 
 ## Inputs
 
diff --git a/built-in-nodes/KlingStartEndFrameNode.mdx b/built-in-nodes/KlingStartEndFrameNode.mdx
index 3f7a0e00f..cbfd41aea 100644
--- a/built-in-nodes/KlingStartEndFrameNode.mdx
+++ b/built-in-nodes/KlingStartEndFrameNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: Legacy model options using `kling-v1-5` and `kling-v1-6` are scheduled for retirement on **September 15, 2026**. Use [`api_kling_v3_flf2v`](/tutorials/partner-nodes/kling/kling-3-0#legacy-kling-models-eol) or Kling O3 FLF2V instead.
-</Warning>
-
-This node creates a video sequence that transitions between your provided start and end images.
+This node creates a video sequence that transitions between your provided start and end images. It generates all the frames in between to produce a smooth transformation from the first frame to the last frame. This node calls the image-to-video API but only supports the input options that work with the `image_tail` request field.
 
 ## Inputs
 
diff --git a/built-in-nodes/KlingVirtualTryOnNode.mdx b/built-in-nodes/KlingVirtualTryOnNode.mdx
index 20bf69de5..a586c88fc 100644
--- a/built-in-nodes/KlingVirtualTryOnNode.mdx
+++ b/built-in-nodes/KlingVirtualTryOnNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: The Kolors Virtual Try-On API (`kolors-virtual-try-on`, `kolors-virtual-try-on-v1-5`) is scheduled for retirement on **September 15, 2026**. Use [Flux Virtual Try-On](/built-in-nodes/FluxVTONode) instead.
-</Warning>
-
-Kling Virtual Try On Node. Input a human image and a cloth image to try on the cloth on the human.
+Kling Virtual Try On Node. Input a human image and a cloth image to try on the cloth on the human. You can merge multiple clothing item pictures into one image with a white background.
 
 ## Inputs
 
diff --git a/built-in-nodes/Load3DAdvanced.mdx b/built-in-nodes/Load3DAdvanced.mdx
index 5c1aec8b3..f5f6ec661 100644
--- a/built-in-nodes/Load3DAdvanced.mdx
+++ b/built-in-nodes/Load3DAdvanced.mdx
@@ -6,31 +6,32 @@ icon: "circle"
 mode: wide
 ---
 
-The Load 3D (Advanced) node loads a 3D model file from your ComfyUI input directory and provides the model data along with camera and viewport information. It supports common 3D file formats and allows you to specify the output image dimensions for rendering.
+The Load 3D (Advanced) node loads a 3D model file from your ComfyUI input directory and provides the model data along with camera and viewport information. It supports common 3D file formats and allows you to specify the output image dimensions.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `model_file` | The 3D model file to load. Select "none" to skip loading a model. | STRING | Yes | `"none"`<br />List of available 3D files in the input/3d directory |
+| `model_file` | The 3D model file to load. Select "none" to skip loading a model. | STRING | Yes | `"none"`<br />List of available 3D files in the `input/3d` directory |
 | `viewport_state` | The current viewport state containing camera and model information from the 3D viewer. | LOAD3D | Yes | - |
-| `width` | The width of the output image in pixels (default: 1024) | INT | Yes | Min: 1<br />Max: 4096<br />Step: 1 |
-| `height` | The height of the output image in pixels (default: 1024) | INT | Yes | Min: 1<br />Max: 4096<br />Step: 1 |
+| `width` | The width of the output image in pixels (default: 1024). | INT | Yes | Min: 1<br />Max: 4096<br />Step: 1 |
+| `height` | The height of the output image in pixels (default: 1024). | INT | Yes | Min: 1<br />Max: 4096<br />Step: 1 |
 
 **Notes on Parameters:**
 - The `model_file` parameter only shows files with the following extensions: .gltf, .glb, .obj, .fbx, .stl
 - Files must be placed in the `input/3d` directory of your ComfyUI installation
 - If `model_file` is set to "none", no 3D model data will be loaded (the output `model_3d` will be empty)
+- If `model_file` is set to a file that does not exist, the node returns a validation error: "Invalid 3D model file: {model_file}"
 
 ## Outputs
 
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
-| `model_3d` | The loaded 3D model data, or empty if no model file was selected | FILE3DANY |
-| `model_3d_info` | Information about the loaded 3D model from the viewport state | LOAD3DMODELINFO |
-| `camera_info` | Camera information from the viewport state | LOAD3DCAMERA |
-| `width` | The specified output image width | INT |
-| `height` | The specified output image height | INT |
+| `model_3d` | The loaded 3D model data, or empty if no model file was selected. | FILE3DANY |
+| `model_3d_info` | Information about the loaded 3D model from the viewport state. | LOAD3DMODELINFO |
+| `camera_info` | Camera information from the viewport state. | LOAD3DCAMERA |
+| `width` | The specified output image width. | INT |
+| `height` | The specified output image height. | INT |
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/Load3DAdvanced/en.md)
 
diff --git a/built-in-nodes/LoadDA3Model.mdx b/built-in-nodes/LoadDA3Model.mdx
index 0d5f51ee6..cf8cffe92 100644
--- a/built-in-nodes/LoadDA3Model.mdx
+++ b/built-in-nodes/LoadDA3Model.mdx
@@ -6,14 +6,14 @@ icon: "circle"
 mode: wide
 ---
 
-The **Load Depth Anything 3** node loads a Depth Anything 3 model from a file, preparing it for depth estimation tasks. You can select the model file and optionally choose the numerical precision (data type) for the model’s weights.
+The **Load Depth Anything 3** node loads a Depth Anything 3 model from a file, preparing it for depth estimation tasks. You can select the model file and optionally choose the numerical precision (data type) for the model's weights.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `model_name` | The name of the Depth Anything 3 model file to load. | STRING | Yes | List of available model files in the `geometry_estimation` folder |
-| `weight_dtype` | The numerical precision (data type) for the model weights. The “default” option uses the model’s original precision. (default: “default”) | STRING | No | `"default"`<br />`"fp16"`<br />`"bf16"`<br />`"fp32"` |
+| `weight_dtype` | The numerical precision (data type) for the model weights. "default" uses the model's original precision, "fp16" uses 16-bit floating point, "bf16" uses bfloat16, and "fp32" uses 32-bit floating point. (default: "default") | STRING | No | `"default"`<br />`"fp16"`<br />`"bf16"`<br />`"fp32"` |
 
 ## Outputs
 
diff --git a/built-in-nodes/LoadVideoDataSetFromFolder.mdx b/built-in-nodes/LoadVideoDataSetFromFolder.mdx
index c89a816a8..109026100 100644
--- a/built-in-nodes/LoadVideoDataSetFromFolder.mdx
+++ b/built-in-nodes/LoadVideoDataSetFromFolder.mdx
@@ -6,19 +6,21 @@ icon: "circle"
 mode: wide
 ---
 
-Load a dataset of videos from a specified folder within the ComfyUI input directory. The node scans the folder for supported video files and returns lazy references — actual frames are decoded only when needed downstream.
+Loads all supported video files from a selected folder inside the ComfyUI input directory and returns them as a list of video references. This node returns lazy video references, so frames are decoded only when another node actually needs them.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `folder` | The folder containing video files. Select from available subfolders inside the ComfyUI input directory. | STRING | Yes | *(populated from input subfolders)* |
+| `folder` | The folder containing video files. Select from available subfolders inside the ComfyUI input directory. | STRING | Yes | All subfolders available in the ComfyUI input directory |
+
+**Note:** The selected folder must contain at least one supported video file. Supported extensions are MP4, AVI, MOV, WEBM, MKV, and FLV. If no supported video files are found, the node raises an error.
 
 ## Outputs
 
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
-| `videos` | A list of lazy video references (one per file). Video frames are decoded only when the output is consumed by another node. | VIDEO |
+| `videos` | A list of lazy video references, one for each video file in the selected folder. Video frames are decoded only when the output is consumed by another node. | VIDEO (list) |
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LoadVideoDataSetFromFolder/en.md)
 
diff --git a/built-in-nodes/LoadVideoTextDataSetFromFolder.mdx b/built-in-nodes/LoadVideoTextDataSetFromFolder.mdx
index 9f919a755..e4e5d689a 100644
--- a/built-in-nodes/LoadVideoTextDataSetFromFolder.mdx
+++ b/built-in-nodes/LoadVideoTextDataSetFromFolder.mdx
@@ -6,20 +6,22 @@ icon: "circle"
 mode: wide
 ---
 
-This node loads a dataset of video files and their corresponding text captions from a specified subfolder inside the ComfyUI input directory. It returns two lists: lazy video references (frames are decoded only when needed downstream) and their associated captions. The node supports common video formats such as MP4, AVI, MOV, WEBM, MKV, and FLV, and can also handle nested folder structures with repeat‑count prefixes (e.g. `5_classname/`) used by tools like kohya‑ss/sd‑scripts.
+This node loads video files and their associated text captions from a selected subfolder in the ComfyUI input directory, and returns them as two lists: videos and captions. The video entries are lazy references, so frames are decoded only when a downstream node needs them. Supported formats are MP4, AVI, MOV, WEBM, MKV, and FLV. Nested folders with a repeat-count prefix (for example `5_classname/`, as used by tools like kohya-ss/sd-scripts) are also supported.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `folder` | The subfolder containing video files and `.txt` caption files. Select from the available subfolders in the ComfyUI input directory. | STRING | Yes | Combo: dynamic list of all subdirectories inside the ComfyUI input folder |
+| `folder` | The folder containing video files and .txt captions. | STRING | Yes | Combo: dynamic list of all subfolders inside the ComfyUI input directory |
+
+If the selected folder contains no files with a supported video extension, the node raises an error. For nested folders whose name starts with a number followed by an underscore (for example `5_classname`), each video inside that folder is included in the dataset the number of times given by that prefix.
 
 ## Outputs
 
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
-| `videos` | Lazy references to the loaded video files. Frames are decoded only when connected to a downstream node that processes them. Each element corresponds to a video from the input folder. | VIDEO (list) |
-| `texts` | List of text captions, one per video. If a video does not have a matching `.txt` file, its caption is an empty string. | STRING (list) |
+| `videos` | Lazy video references; frames are decoded only when needed downstream. One entry per video file found in the folder. | VIDEO (list) |
+| `texts` | List of text captions. One caption per video; if a video has no matching `.txt` file, its caption is an empty string. | STRING (list) |
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LoadVideoTextDataSetFromFolder/en.md)
 
diff --git a/built-in-nodes/LumaRay32ExtendVideoNode.mdx b/built-in-nodes/LumaRay32ExtendVideoNode.mdx
index 9b69ab80e..8a4e9fb3f 100644
--- a/built-in-nodes/LumaRay32ExtendVideoNode.mdx
+++ b/built-in-nodes/LumaRay32ExtendVideoNode.mdx
@@ -6,20 +6,20 @@ icon: "circle"
 mode: wide
 ---
 
-Luma Ray 3.2 Extend Video extends a previous Luma Ray 3.2 video generation by adding new content either after it (forward extension) or before it (backward extension). Connect the generation ID output from a prior Luma Ray 3.2 node to create a seamless 5-second extension of your video.
+Luma Ray 3.2 Extend Video continues a previous Luma Ray 3.2 video generation by creating a new 5-second segment either after the original clip (forward) or before it (backward). Connect the `generation_id` output from an earlier Luma Ray 3.2 node to use that clip as the starting frame (forward) or ending frame (backward) of the extension.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `source_generation_id` | Generation ID of the prior Ray 3.2 video to extend. Connect the generation_id output of another Luma Ray 3.2 node. | STRING | Yes | – |
-| `direction` | Forward continues after the prior clip; backward is prepended before it. When "Forward (continue after)" is selected, you can optionally enable loop mode. | COMBO | Yes | "Forward (continue after)"<br />"Backward (lead-in before)" |
-| `loop` | Loop the extended video seamlessly (forward extend only). | BOOLEAN | No | True<br />False |
-| `prompt` | Text prompt for the new content. | STRING | Yes | – |
-| `resolution` | Output resolution for the extended video segment. | COMBO | Yes | "540p"<br />"720p"<br />"1080p" |
+| `source_generation_id` | Generation ID of the prior Ray 3.2 video to extend. Connect the `generation_id` output of another Luma Ray 3.2 node. This value is required and must not be empty. | STRING | Yes | – |
+| `direction` | Forward continues after the prior clip; backward is prepended before it. Selecting "Forward (continue after)" also adds the `loop` option. | COMBO | Yes | "Forward (continue after)"<br />"Backward (lead-in before)" |
+| `loop` | Loop the extended video seamlessly (forward extend only). Only available when `direction` is "Forward (continue after)". Default: False. | BOOLEAN | No | True<br />False |
+| `prompt` | Text prompt for the new content. Must be between 1 and 6000 characters. | STRING | Yes | – |
+| `resolution` | Output resolution for the extended video segment. Default: "720p". | COMBO | Yes | "540p"<br />"720p"<br />"1080p" |
 | `seed` | Random seed for reproducible generation results. | INT | Yes | – |
 
-**Note:** The `loop` parameter is only available when `direction` is set to "Forward (continue after)". When using "Backward (lead-in before)", the loop option is not available. The `prompt` must be between 1 and 6000 characters. The `source_generation_id` is required and must be connected from a prior Luma Ray 3.2 node's generation_id output.
+**Note:** The `loop` parameter is only available when `direction` is set to "Forward (continue after)". When using "Backward (lead-in before)", the `loop` option is not available. The `prompt` must be between 1 and 6000 characters. The `source_generation_id` is required and must be connected from a prior Luma Ray 3.2 node's `generation_id` output.
 
 ## Outputs
 
diff --git a/built-in-nodes/LumaRay32KeyframeNode.mdx b/built-in-nodes/LumaRay32KeyframeNode.mdx
index 21e73caac..6bef76619 100644
--- a/built-in-nodes/LumaRay32KeyframeNode.mdx
+++ b/built-in-nodes/LumaRay32KeyframeNode.mdx
@@ -6,6 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
+```markdown
 This node anchors a guide image to a specific position on the Luma Ray 3.2 output video timeline. Connect this node to the "keyframes" input of the Luma Ray 3.2 Keyframes to Video node, and chain several keyframes together by connecting the optional "keyframes" input.
 
 ## Inputs
@@ -25,6 +26,7 @@ When "Absolute time (seconds)" is selected for the `position` parameter, you can
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
 | `keyframes` | A keyframe chain that includes the new keyframe combined with any optional earlier keyframes. | LUMA_RAY32_KEYFRAME |
+```
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LumaRay32KeyframeNode/en.md)
 
diff --git a/built-in-nodes/LumaRay32KeyframesToVideoNode.mdx b/built-in-nodes/LumaRay32KeyframesToVideoNode.mdx
index a19b43974..45bf31db7 100644
--- a/built-in-nodes/LumaRay32KeyframesToVideoNode.mdx
+++ b/built-in-nodes/LumaRay32KeyframesToVideoNode.mdx
@@ -12,7 +12,7 @@ This node generates a video that interpolates through a sequence of guide images
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `prompt` | Text prompt for the video generation. | STRING | Yes | 1 to 6000 characters |
+| `prompt` | Text prompt for the video generation (default: empty string). | STRING | Yes | 1 to 6000 characters |
 | `resolution` | The output resolution of the generated video (default: "720p"). | STRING | Yes | `"360p"`<br />`"540p"`<br />`"720p"`<br />`"1080p"` |
 | `duration` | The duration of the generated video (default: "5s"). | STRING | Yes | `"5s"`<br />`"10s"` |
 | `seed` | Seed for random number generation to control reproducibility. | INT | Yes | 0 to 4294967295 |
diff --git a/built-in-nodes/LumaRay32TextToVideoNode.mdx b/built-in-nodes/LumaRay32TextToVideoNode.mdx
index d6f5449ed..6eca03291 100644
--- a/built-in-nodes/LumaRay32TextToVideoNode.mdx
+++ b/built-in-nodes/LumaRay32TextToVideoNode.mdx
@@ -12,14 +12,14 @@ This node generates a video from a text prompt using Luma's Ray 3.2 model. It se
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `prompt` | Text prompt for the video generation. | STRING | Yes | 1-6000 characters |
+| `prompt` | Text prompt for the video generation. | STRING | Yes | 1 to 6000 characters |
 | `aspect_ratio` | The aspect ratio of the generated video. | STRING | Yes | `"16:9"`<br />`"9:16"`<br />`"1:1"`<br />`"4:3"`<br />`"3:4"`<br />`"21:9"` |
 | `resolution` | The output resolution of the video (default: "720p"). | STRING | Yes | `"360p"`<br />`"540p"`<br />`"720p"`<br />`"1080p"` |
 | `duration` | The duration of the generated video. | STRING | Yes | `"5s"`<br />`"10s"` |
 | `loop` | Make the video loop seamlessly. Only available with 5s duration. | BOOLEAN | No | True/False (default: False) |
 | `seed` | Seed for reproducible generation. | INT | No | 0 to 2147483647 |
 
-**Note:** The `loop` parameter can only be enabled when `duration` is set to "5s". If you select "10s" duration and enable looping, the node will return an error.
+**Note:** The `loop` parameter can only be enabled when `duration` is set to "5s". If you set `duration` to "10s" and enable `loop`, the node returns an error.
 
 ## Outputs
 
diff --git a/built-in-nodes/LumaRay32VideoEditNode.mdx b/built-in-nodes/LumaRay32VideoEditNode.mdx
index fd2cbdfde..5952b4a3f 100644
--- a/built-in-nodes/LumaRay32VideoEditNode.mdx
+++ b/built-in-nodes/LumaRay32VideoEditNode.mdx
@@ -16,7 +16,7 @@ This node re-renders an existing video under a new prompt using Luma Ray 3.2, al
 |-----------|-------------|-----------|----------|-------|
 | `video` | Source video to edit. Up to 18 seconds. | VIDEO | Yes | - |
 | `prompt` | Describes the desired edit. | STRING | Yes | - |
-| `resolution` | The output resolution for the edited video. | COMBO | Yes | `"360p"`<br />`"540p"`<br />`"720p"`<br />`"1080p"` |
+| `resolution` | The output resolution for the edited video. (default: "720p") | COMBO | Yes | `"360p"`<br />`"540p"`<br />`"720p"`<br />`"1080p"` |
 | `strength` | How strongly to preserve vs. reimagine the source. "auto" lets Ray 3.2 choose; adhere_* preserves the most, flex_* is balanced, reimagine_* changes the most. (default: "auto") | COMBO | Yes | `"auto"`<br />`"adhere_1"`<br />`"adhere_2"`<br />`"adhere_3"`<br />`"flex_1"`<br />`"flex_2"`<br />`"flex_3"`<br />`"reimagine_1"`<br />`"reimagine_2"`<br />`"reimagine_3"` |
 | `seed` | Seed for reproducibility. | INT | Yes | - |
 
diff --git a/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx b/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
new file mode 100644
index 000000000..dbd10caf4
--- /dev/null
+++ b/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
@@ -0,0 +1,36 @@
+---
+title: "MinimaxHailuo03FirstLastFrameNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03FirstLastFrameNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03FirstLastFrameNode"
+icon: "circle"
+mode: wide
+---
+
+This node generates a video from a first frame image and an optional last frame image using the MiniMax H3 model. The video follows the aspect ratio of the supplied images and, when a last frame is provided, animates from the first frame toward the last frame.
+
+## Inputs
+
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `model` | Model to use for video generation. This combo includes the model choice ("MiniMax H3"), a text prompt describing the video to generate, the output resolution, and the video duration. The prompt must contain at least one non-whitespace character. | COMBO | Yes | "MiniMax H3" |
+| `first_frame` | First frame image for the video. The aspect ratio of the generated video follows this image. Must be at least 256x256 pixels with a width-to-height aspect ratio between 0.4 and 2.5. | IMAGE | Yes | - |
+| `last_frame` | Optional last frame image for the video. When provided, the video starts from the first frame and ends at this image. Must meet the same size and aspect ratio requirements as `first_frame`. | IMAGE | No | - |
+| `seed` | Random seed. The same request with the same seed gives similar, but not guaranteed identical, results. Includes a "control after generate" option to randomize after each generation. Default: 42. | INT | Yes | 0 to 4294967295 |
+| `watermark` | Whether to add an AIGC watermark to the video. This is an advanced parameter. Default: False. | BOOLEAN | Yes | True<br />False |
+
+**Note on constraints:**
+- The text prompt inside the `model` combo cannot be empty; whitespace-only prompts are rejected.
+- Any provided frame image (`first_frame` and, if used, `last_frame`) must be at least 256 pixels wide and 256 pixels tall, with a width-to-height aspect ratio between 0.4 and 2.5 (approximately 2:5 to 5:2).
+- `last_frame` is optional. When omitted, the video is generated from the first frame only.
+- The aspect ratio of the output video is determined by the supplied images, not by a separate ratio setting.
+
+## Outputs
+
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `video` | The generated video created from the first frame and optional last frame using the MiniMax H3 model. | VIDEO |
+
+> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03FirstLastFrameNode/en.md)
+
+---
+**Source fingerprint (SHA-256):** `f4cb9217eb346019680c64b30c1beacce16f0050616b7b76265edc5840f6b21e`
diff --git a/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx b/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
new file mode 100644
index 000000000..567c71d1f
--- /dev/null
+++ b/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
@@ -0,0 +1,42 @@
+---
+title: "MinimaxHailuo03ReferenceNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03ReferenceNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03ReferenceNode"
+icon: "circle"
+mode: wide
+---
+
+This node generates a video using the MiniMax H3 model, using reference images, videos, and audio to condition the result. References are referred to in the prompt by their connection order: "Image 1", "Image 2", "Video 1", "Audio 1", and so on.
+
+## Inputs
+
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `model` | Model to use for video generation (default: "MiniMax H3"). Selecting "MiniMax H3" provides the `prompt`, `duration`, `resolution`, `ratio`, `reference_images`, `reference_videos`, and `reference_audios` settings below. | STRING | Yes | "MiniMax H3" |
+| `prompt` | Text description of the video to generate. Reference media can be referred to by their order, for example "Image 1", "Image 2", "Video 1", or "Audio 1". | STRING | Yes | Min length: 1 character |
+| `duration` | Duration of the generated video in seconds. | INT | Yes | Multiple options available |
+| `resolution` | Output resolution of the generated video. | STRING | Yes | Multiple options available |
+| `ratio` | Aspect ratio of the generated video. | STRING | Yes | Multiple options available |
+| `reference_images` | Subject or style reference images, referred to in the prompt as "Image 1".."Image 9" in connection order. Up to 9 images. | IMAGE | No | 0 to 9 images |
+| `reference_videos` | Motion or scene reference videos, referred to in the prompt as "Video 1".."Video 3" in connection order. Up to 3 videos, 2-15 seconds each, 15 seconds in total. | VIDEO | No | 0 to 3 videos |
+| `reference_audios` | Audio references, referred to in the prompt as "Audio 1".."Audio 3" in connection order. Up to 3 clips, 2-15 seconds each, 15 seconds in total. Cannot be used without a reference image or video. | AUDIO | No | 0 to 3 clips |
+| `seed` | Random seed. The same request with the same seed gives similar, but not guaranteed identical, results (default: 42). | INT | Yes | 0 to 4294967295 |
+| `watermark` | Whether to add an AIGC watermark to the video (default: false). | BOOLEAN | No | true<br />false |
+
+### Parameter Constraints
+
+- At least one reference image or one reference video is required. Reference audio alone is not accepted.
+- Each reference image must have an aspect ratio between approximately 0.4 and 2.5 (2:5 to 5:2) and a minimum width and height of 256 pixels.
+- Each reference video must be between 2 and 15 seconds long with a frame rate between 23.976 and 60 FPS. The total duration of all reference videos cannot exceed 15 seconds.
+- Each reference audio clip must be between 2 and 15 seconds long. The total duration of all reference audio clips cannot exceed 15 seconds.
+
+## Outputs
+
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `video` | The generated video. | VIDEO |
+
+> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03ReferenceNode/en.md)
+
+---
+**Source fingerprint (SHA-256):** `beca020333a544188e6c21829eb8e63415aa5299efc676438e85662a5f08660d`
diff --git a/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx b/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
new file mode 100644
index 000000000..a6d088d4e
--- /dev/null
+++ b/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
@@ -0,0 +1,30 @@
+---
+title: "MinimaxHailuo03TextToVideoNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03TextToVideoNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03TextToVideoNode"
+icon: "circle"
+mode: wide
+---
+
+This node generates a video from a text prompt using the MiniMax H3 model. It sends the text along with video settings such as resolution, duration, and aspect ratio to the MiniMax API, and returns the resulting video as its output.
+
+## Inputs
+
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `model` | Model to use for video generation. (default: "MiniMax H3"). This selection also includes the text prompt, resolution, duration, and aspect ratio settings for the generated video. | COMBO | Yes | `"MiniMax H3"` |
+| `seed` | Random seed. The same request with the same seed gives similar, but not guaranteed identical, results. (default: 42) | INT | Yes | 0 to 4294967295 |
+| `watermark` | Whether to add an AIGC watermark to the video. (default: false) | BOOLEAN | No | true<br />false |
+
+Note: The text prompt included in the `model` option must contain at least one non-whitespace character. The estimated price shown for this node is calculated from the selected video duration.
+
+## Outputs
+
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `VIDEO` | The generated video created from the provided text prompt. | VIDEO |
+
+> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03TextToVideoNode/en.md)
+
+---
+**Source fingerprint (SHA-256):** `9478576dd02ed407a39c95c7227eb8e1482db8b77adc814691fbd807e4cc2893`
diff --git a/built-in-nodes/OpenAIGPTImageNodeV2.mdx b/built-in-nodes/OpenAIGPTImageNodeV2.mdx
index f2969cb91..e275b4aae 100644
--- a/built-in-nodes/OpenAIGPTImageNodeV2.mdx
+++ b/built-in-nodes/OpenAIGPTImageNodeV2.mdx
@@ -28,7 +28,7 @@ This node generates images using OpenAI's GPT Image API. It supports multiple mo
 
 **Parameter Constraints and Limitations:**
 
-- When using `gpt-image-2` with a `model.size` of "Custom", the `custom_width` and `custom_height` must be multiples of 16, the maximum edge must not exceed 3840, the aspect ratio must not exceed 3:1, and the total pixel count must be between 655,360 and 8,294,400.
+- When using `gpt-image-2` with a `model.size` of "Custom", the `custom_width` and `custom_height` must be multiples of 16, the maximum edge must be `<= 3840`, the aspect ratio must not exceed 3:1, and the total pixel count must be between 655,360 and 8,294,400.
 - If a `mask` is provided, an input image (`model.images`) is required. A mask cannot be used without an input image.
 - A mask cannot be used with multiple input images.
 - When a mask is provided, the mask dimensions must match the input image dimensions.
diff --git a/built-in-nodes/RunwayAleph2KeyframeNode.mdx b/built-in-nodes/RunwayAleph2KeyframeNode.mdx
index 17dacf826..f84ededb7 100644
--- a/built-in-nodes/RunwayAleph2KeyframeNode.mdx
+++ b/built-in-nodes/RunwayAleph2KeyframeNode.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-This node anchors a guidance image to a specific moment of your input video, so the Aleph2 model steers the edit at that point in your footage. Connect this node to the "keyframes" input of the Runway Aleph2 Video to Video node, and chain several together (up to 5) via the optional "keyframes" input.
+Runway Aleph2 Keyframe node anchors a guidance image to a specific moment of your input video, so the Aleph2 model steers the edit at that point in your footage. Connect this node to the "keyframes" input of the Runway Aleph2 Video to Video node, and chain several together (up to 5) via the optional "keyframes" input.
 
 ## Inputs
 
diff --git a/built-in-nodes/RunwayAleph2PromptImageNode.mdx b/built-in-nodes/RunwayAleph2PromptImageNode.mdx
index 60ff6c7f4..1a269e050 100644
--- a/built-in-nodes/RunwayAleph2PromptImageNode.mdx
+++ b/built-in-nodes/RunwayAleph2PromptImageNode.mdx
@@ -13,7 +13,7 @@ This node anchors a guidance image to a specific moment in the output video, con
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `image` | The guidance image to place at the chosen moment of the output video. | IMAGE | Yes | - |
-| `position` | How to place this image on the output video's timeline. Choose between absolute timing (seconds from start) or fractional timing (percentage of video duration). Selecting an option reveals the corresponding sub-parameter. | COMBO | Yes | `Absolute (seconds)`<br />`Fraction (0.0 to 1.0)` |
+| `position` | How to place this image on the output video's timeline. Choose between absolute timing (seconds from start) or fractional timing (fraction of video duration). Selecting an option reveals the corresponding sub-parameter. | COMBO | Yes | `Absolute (seconds)`<br />`Fraction (0.0 to 1.0)` |
 | `prompt_images` | Optional earlier prompt images to chain with this one. Connect the output of another Runway Aleph2 Prompt Image node here to build a chain of up to 5 guidance images. | PROMPT_IMAGE_CHAIN | No | - |
 
 **Position mode details:**
diff --git a/built-in-nodes/RunwayFirstLastFrameNode.mdx b/built-in-nodes/RunwayFirstLastFrameNode.mdx
index cf020b574..cc472f963 100644
--- a/built-in-nodes/RunwayFirstLastFrameNode.mdx
+++ b/built-in-nodes/RunwayFirstLastFrameNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: This node reached end of life on **July 30, 2026** when Runway retired Gen-3 Alpha Turbo. For first-last-frame video, use [Kling 3.0 FLF2V](/tutorials/partner-nodes/kling/kling-3-0) or other supported FLF2V templates.
-</Warning>
-
-The Runway First-Last-Frame to Video node generates videos by uploading first and last keyframes along with a text prompt.
+The Runway First-Last-Frame to Video node generates videos by uploading first and last keyframes along with a text prompt. It creates smooth transitions between the provided start and end frames using Runway's Gen-3 model. This is particularly useful for complex transitions where the end frame differs significantly from the start frame.
 
 ## Inputs
 
diff --git a/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx b/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
index 31b966a86..6a41652a4 100644
--- a/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
+++ b/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
@@ -6,12 +6,6 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **End of life (EOL)**: This node reached end of life on **July 30, 2026** when Runway retired Gen-3 Alpha Turbo. Use [Runway Gen4 Turbo](/built-in-nodes/RunwayImageToVideoNodeGen4) or the [Gen4 Turbo workflow](/tutorials/partner-nodes/runway/video-generation) instead.
-</Warning>
-
 The Runway Image to Video (Gen3a Turbo) node generates a video from a single starting frame using Runway's Gen3a Turbo model. It takes a text prompt and an initial image frame, then creates a video sequence based on the specified duration and aspect ratio. This node connects to Runway's API to process the generation remotely.
 
 ## Inputs
diff --git a/built-in-nodes/SCAIL2ColoredMask.mdx b/built-in-nodes/SCAIL2ColoredMask.mdx
index 798f71447..51e558dc7 100644
--- a/built-in-nodes/SCAIL2ColoredMask.mdx
+++ b/built-in-nodes/SCAIL2ColoredMask.mdx
@@ -18,6 +18,8 @@ This node renders SAM3 tracking data into colored masks that are consumed by the
 | `sort_by` | Order in which palette colors are assigned to the tracked objects (applied to both reference and pose video so each identity keeps the same color). Objects that appear in earlier frames always come first; within a frame, left_to_right = leftmost object (by centroid at first appearance) gets the first color, area = biggest object (by mask area at first appearance) gets the first color; none = keep SAM3's order. (default: "left_to_right") | COMBO | Yes | `"none"`<br />`"left_to_right"`<br />`"area"` |
 | `replacement_mode` | False = Animation Mode (pose_video_mask has black background, reference_image_mask has white background). True = Replacement Mode (pose_video_mask has white background, reference_image_mask has black background). (default: False) | BOOLEAN | Yes | False<br />True |
 
+Note: `object_indices` only accepts comma-separated digits; non-numeric entries and out-of-range indices are ignored. When `ref_track_data` is not provided, the `reference_image_mask` output is a solid fill using the reference background color (white in Animation Mode, black in Replacement Mode).
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/Save3DAdvanced.mdx b/built-in-nodes/Save3DAdvanced.mdx
index 1f3b51fcf..a8b2c195e 100644
--- a/built-in-nodes/Save3DAdvanced.mdx
+++ b/built-in-nodes/Save3DAdvanced.mdx
@@ -12,7 +12,7 @@ This node saves a 3D model to a file in the ComfyUI output directory, with advan
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `model_3d` | 3D model file from an upstream 3D node. | FILE3D | Yes | GLB, GLTF, FBX, OBJ, STL, USDZ, Any |
+| `model_3d` | 3D model file from an upstream 3D node. | FILE3D | Yes | GLB<br />GLTF<br />FBX<br />OBJ<br />STL<br />USDZ<br />Any |
 | `filename_prefix` | Prefix for the saved file name (default: "3d/ComfyUI"). | STRING | Yes | Free text |
 | `viewport_state` | Viewport state from a Load 3D node, containing camera and model information. | LOAD3D | Yes | - |
 | `model_3d_info` | Optional 3D model information to override the viewport state. | LOAD3DMODELINFO | No | - |
@@ -20,6 +20,8 @@ This node saves a 3D model to a file in the ComfyUI output directory, with advan
 | `width` | Width of the output preview in pixels (default: 1024). | INT | Yes | 1 to 4096 |
 | `height` | Height of the output preview in pixels (default: 1024). | INT | Yes | 1 to 4096 |
 
+Note: When `model_3d_info` or `camera_info` are not connected, the node falls back to the values stored in `viewport_state`.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/SaveGaussianSplat.mdx b/built-in-nodes/SaveGaussianSplat.mdx
index 3ff1a2d2a..92887cba6 100644
--- a/built-in-nodes/SaveGaussianSplat.mdx
+++ b/built-in-nodes/SaveGaussianSplat.mdx
@@ -6,11 +6,6 @@ icon: "circle"
 mode: wide
 ---
 
-The existing documentation is accurate and aligns with the source code. No updates are necessary.
-
-```markdown
-# SaveGaussianSplat
-
 This node saves a Gaussian splat 3D file to the output directory. It handles the file saving process and provides preview data for the 3D viewport.
 
 ## Inputs
@@ -25,16 +20,17 @@ This node saves a Gaussian splat 3D file to the output directory. It handles the
 | `width` | The width of the preview (default: 1024). | INT | Yes | 1 to 4096 |
 | `height` | The height of the preview (default: 1024). | INT | Yes | 1 to 4096 |
 
+Note: If `camera_info` or `model_3d_info` are not provided, the node uses the corresponding values from `viewport_state` when available.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
 |-------------|-------------|-----------|
-| `model_3d` | The saved Gaussian splat 3D file. | FILE3D |
+| `model_3d` | The Gaussian splat 3D file that was saved to the output directory. | FILE3D |
 | `model_3d_info` | Model 3D information for the viewport. | LOAD3DMODELINFO |
 | `camera_info` | Camera information for the viewport preview. | LOAD3DCAMERA |
 | `width` | The width of the preview. | INT |
 | `height` | The height of the preview. | INT |
-```
 
 > This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/SaveGaussianSplat/en.md)
 
diff --git a/built-in-nodes/SavePointCloud.mdx b/built-in-nodes/SavePointCloud.mdx
index 0b7535432..ae2603604 100644
--- a/built-in-nodes/SavePointCloud.mdx
+++ b/built-in-nodes/SavePointCloud.mdx
@@ -20,6 +20,8 @@ The Save Point Cloud node saves a 3D point cloud file to the output directory an
 | `width` | Width of the preview display in pixels (default: 1024) | INT | Yes | 1 to 4096 |
 | `height` | Height of the preview display in pixels (default: 1024) | INT | Yes | 1 to 4096 |
 
+The saved file is written to the output directory using the `filename_prefix` with a 5-digit counter appended (for example, `3d/ComfyUI_00001.ply`). When the optional `camera_info` input is not connected, the camera information is taken from `viewport_state`; similarly, when `model_3d_info` is not connected, the model information is taken from `viewport_state` (or an empty list if none is available).
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/SaveText.mdx b/built-in-nodes/SaveText.mdx
index e5f518e8d..8b53a00bc 100644
--- a/built-in-nodes/SaveText.mdx
+++ b/built-in-nodes/SaveText.mdx
@@ -12,9 +12,9 @@ The Save Text node writes text content to a file in the output directory. It sup
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `text` | The text content to save to a file | STRING | Yes | - |
-| `filename_prefix` | Prefix for the output filename (default: "ComfyUI") | STRING | No | - |
-| `format` | The file format to save the text as (default: "txt") | STRING | No | `"txt"`<br />`"md"`<br />`"json"` |
+| `text` | The text content to save to a file. This input must be connected from another node. | STRING | Yes | - |
+| `filename_prefix` | Prefix for the output filename. A 5-digit counter is appended to prevent overwriting existing files (default: "ComfyUI"). | STRING | No | - |
+| `format` | The file format to save the text as (default: "txt"). When "json" is selected, valid JSON text is pretty-printed with 2-space indentation; otherwise, the text is saved as-is. | COMBO | No | `"txt"`<br />`"md"`<br />`"json"` |
 
 ## Outputs
 
diff --git a/built-in-nodes/SeedNode.mdx b/built-in-nodes/SeedNode.mdx
index 33ff4c40d..bebe133f8 100644
--- a/built-in-nodes/SeedNode.mdx
+++ b/built-in-nodes/SeedNode.mdx
@@ -6,13 +6,13 @@ icon: "circle"
 mode: wide
 ---
 
-The Seed node generates a fixed or random integer value. It is commonly used to control the reproducibility of random operations in other nodes by providing a consistent starting point for their random number generation.
+The Seed node provides an integer value that can be used as a seed to control the reproducibility of random operations in other nodes. By supplying a consistent starting value, it helps keep generated results repeatable when needed.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `seed` | The seed value to use. The control after generate option determines whether the value stays fixed or changes after each generation. | INT | Yes | 0 to 9223372036854775807 |
+| `seed` | The seed value to use. The control after generate option determines whether the value stays fixed or changes after each generation; in this node it is set to fixed. | INT | Yes | 0 to 9223372036854775807 |
 
 ## Outputs
 
diff --git a/built-in-nodes/SeedVR2TemporalChunk.mdx b/built-in-nodes/SeedVR2TemporalChunk.mdx
index 38aa1b1c2..63de816b8 100644
--- a/built-in-nodes/SeedVR2TemporalChunk.mdx
+++ b/built-in-nodes/SeedVR2TemporalChunk.mdx
@@ -12,17 +12,17 @@ This node splits a SeedVR2 video latent into smaller temporal chunks that can be
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `latent` | The VAE-encoded SeedVR2 latent to split. Must be a 5‑D tensor with the correct channel count. | LATENT | Yes | - |
-| `temporal_overlap` | Latent frames shared between adjacent chunks and crossfaded at merge; 0 means no overlap (default: 0). | INT | No | 0 to 16384 |
+| `latent` | The VAE-encoded SeedVR2 latent to split. Must be a 5-D tensor (B, C, T, H, W) with the expected SeedVR2 latent channel count. | LATENT | Yes | - |
+| `temporal_overlap` | Latent frames shared between adjacent chunks and crossfaded at merge; 0 means no overlap (default: 0). The effective overlap is capped at one less than the chunk's latent frame count. | INT | No | 0 to 16384 |
 | `chunking_mode` | Manual uses `frames_per_chunk` exactly; auto predicts the largest chunk that fits free VRAM. | COMBO | Yes | `"auto"`<br />`"manual"` |
 
 When `chunking_mode` is set to `"manual"`, an additional parameter becomes available:
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
-| `frames_per_chunk` | Pixel frames per temporal chunk. Must be a 4n+1 value (1, 5, 9, 13, 17, 21, ...) (default: 21). | INT | Yes | 1 to 16384 |
+| `frames_per_chunk` | Pixel frames per temporal chunk (4n+1: 1, 5, 9, 13, 17, 21, ...); the UI steps by 4 (default: 21). | INT | Yes | 1 to 16384 |
 
-Note: The `frames_per_chunk` parameter only appears when `chunking_mode` is set to `"manual"`. The value must satisfy the formula `(frames_per_chunk - 1) % 4 == 0`, meaning it must be one of: 1, 5, 9, 13, 17, 21, etc. If the total pixel frames in the latent are equal to or less than the chosen chunk size, the node returns the original latent as a single chunk with zero overlap.
+Note: The `frames_per_chunk` parameter only appears when `chunking_mode` is set to `"manual"`. The value must satisfy the formula `(frames_per_chunk - 1) % 4 == 0`, meaning it must be one of: 1, 5, 9, 13, 17, 21, etc. The input latent must be 5-dimensional and contain the expected SeedVR2 latent channel count; otherwise the node raises an error. If the total pixel frames in the latent are equal to or less than the chosen chunk size (or the auto-computed size), the node returns the original latent as a single chunk with zero overlap.
 
 ## Outputs
 
diff --git a/built-in-nodes/SeedVR2TemporalMerge.mdx b/built-in-nodes/SeedVR2TemporalMerge.mdx
index 743c14baa..21db2efa5 100644
--- a/built-in-nodes/SeedVR2TemporalMerge.mdx
+++ b/built-in-nodes/SeedVR2TemporalMerge.mdx
@@ -6,9 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-Merge SeedVR2 Latents
-
-This node recombines sampled SeedVR2 latent temporal chunks into a single full-length latent. It uses a Hann window crossfade over the overlap region between chunks to create smooth transitions, or performs plain concatenation when no overlap is specified.
+This node recombines sampled SeedVR2 latent temporal chunks into a single full-length latent. When a temporal overlap is specified, it crossfades each overlapping region with a Hann window to create smooth transitions between chunks; when the overlap is 0, it performs plain concatenation.
 
 ## Inputs
 
@@ -17,7 +15,7 @@ This node recombines sampled SeedVR2 latent temporal chunks into a single full-l
 | `latents` | The sampled temporal chunks in sequence order. | LATENT | Yes | List of latents |
 | `temporal_overlap` | The temporal_overlap output of Split SeedVR2 Latent. 0 = plain concatenation. (default: 0) | INT | Yes | 0 to 16384 |
 
-**Note:** The `temporal_overlap` value must be greater than or equal to 0. The final chunk in the sequence may have fewer temporal frames than the other chunks. All chunks must have matching dimensions except for the temporal axis (T), and the first chunk must be 5-dimensional (B, C, T, H, W).
+**Note:** The `temporal_overlap` value must be greater than or equal to 0. All chunks must be 5-dimensional video latents (B, C, T, H, W) and must match in every dimension except the temporal axis (T); only the final chunk may be shorter than the others. If only one chunk is provided, it is returned unchanged.
 
 ## Outputs
 
diff --git a/built-in-nodes/ShuffleVideoDataset.mdx b/built-in-nodes/ShuffleVideoDataset.mdx
index 9e96971bf..911c3b1c1 100644
--- a/built-in-nodes/ShuffleVideoDataset.mdx
+++ b/built-in-nodes/ShuffleVideoDataset.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-This node takes a list of videos and randomly reorders them. It uses a random seed to ensure reproducible shuffling, so the same seed always produces the same output order.
+This node takes a list of videos and randomly reorders them. It uses a random seed to make the shuffle reproducible, so the same seed always produces the same output order.
 
 ## Inputs
 
@@ -15,6 +15,8 @@ This node takes a list of videos and randomly reorders them. It uses a random se
 | `videos` | List of videos to shuffle. | VIDEO | Yes | List of video inputs |
 | `seed` | Random seed for shuffling (default: 0). | INT | No | 0 to 18446744073709551615 |
 
+Note: The seed value is reduced modulo 4294967295 (2^32 - 1) before being used. As a result, seeds that differ by a multiple of 4294967295 produce the same output order.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/ShuffleVideoTextDataset.mdx b/built-in-nodes/ShuffleVideoTextDataset.mdx
index 485549b39..af2d55ae2 100644
--- a/built-in-nodes/ShuffleVideoTextDataset.mdx
+++ b/built-in-nodes/ShuffleVideoTextDataset.mdx
@@ -16,6 +16,8 @@ This node randomly shuffles the order of video-text pairs, keeping each video pa
 | `texts` | List of texts to shuffle. | STRING | Yes | List of text strings |
 | `seed` | Random seed to control the shuffling order (default: 0). | INT | Yes | 0 to 18446744073709551615 |
 
+Note: `videos` and `texts` must have the same length, since the node pairs each video with the text at the same position and preserves those pairings when shuffling.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/SyncLipSyncNode.mdx b/built-in-nodes/SyncLipSyncNode.mdx
index 7a42d3cc4..625f34afc 100644
--- a/built-in-nodes/SyncLipSyncNode.mdx
+++ b/built-in-nodes/SyncLipSyncNode.mdx
@@ -15,22 +15,22 @@ This node re-syncs mouth movement in a video to new speech audio using the sync.
 | `video` | Footage of the speaker to re-sync. Up to 4K (4096x2160); a constant frame rate of 24/25/30 fps works best. | VIDEO | Yes | - |
 | `audio` | Speech audio to sync the mouth to. | AUDIO | Yes | - |
 | `seed` | Seed controls whether the node should re-run; results are non-deterministic regardless of seed (default: 42). | INT | Yes | 0 to 2147483647 |
-| `model` | sync.so generation model. | COMBO | Yes | See below |
+| `model` | sync.so generation model. | COMBO | Yes | "sync-3" |
 
 The `model` parameter is a dynamic combo that includes the following sub-parameters:
 
 | Sub-parameter | Description | Data Type | Required | Range |
 |---------------|-------------|-----------|----------|-------|
-| `sync_mode` | How to handle a duration mismatch between video and audio; this also sets the output length. bounce: video plays forward then backward until the audio ends (output = audio length). loop: video restarts until the audio ends (output = audio length). remap: video is time-stretched to match the audio (output = audio length). cut_off: the longer track is trimmed (output = shorter length). silence: nothing is trimmed; the shorter track is padded (output = longer length). (default: "bounce") | COMBO | Yes | `"bounce"`<br />`"cut_off"`<br />`"loop"`<br />`"silence"`<br />`"remap"` |
-| `speaker_selection` | Which face to lipsync when several people are visible. default: let the model decide. auto-detect: detect and follow the active speaker. coordinates: target the face at pixel (speaker_x, speaker_y) in the frame chosen by speaker_frame. (default: "default") | COMBO | Yes | `"default"`<br />`"auto-detect"`<br />`"coordinates"` |
+| `sync_mode` | How to handle a duration mismatch between video and audio; this also sets the output length. bounce: video plays forward then backward until the audio ends (output = audio length). loop: video restarts until the audio ends (output = audio length). remap: video is time-stretched to match the audio (output = audio length). cut_off: the longer track is trimmed (output = shorter length). silence: nothing is trimmed; the shorter track is padded (output = longer length). (default: "bounce") | COMBO | Yes | "bounce"<br />"cut_off"<br />"loop"<br />"silence"<br />"remap" |
+| `speaker_selection` | Which face to lipsync when several people are visible. default: let the model decide. auto-detect: detect and follow the active speaker. coordinates: target the face at pixel (speaker_x, speaker_y) in the frame chosen by speaker_frame. (default: "default") | COMBO | Yes | "default"<br />"auto-detect"<br />"coordinates" |
 | `speaker_frame` | Video frame used to locate the speaker. Only used when `speaker_selection` is 'coordinates'. (default: 0) | INT | No | 0 to 1,000,000 |
 | `speaker_x` | X pixel coordinate of the speaker's face. Only used when `speaker_selection` is 'coordinates'. (default: 0) | INT | No | 0 to 4096 |
 | `speaker_y` | Y pixel coordinate of the speaker's face. Only used when `speaker_selection` is 'coordinates'. (default: 0) | INT | No | 0 to 4096 |
 
 **Constraints:**
 - Video resolution must not exceed 4K (4096x2160). Videos above this limit will raise an error.
-- Audio duration must not exceed 600 seconds (10 minutes).
-- The `speaker_frame`, `speaker_x`, and `speaker_y` parameters are only used when `speaker_selection` is set to `"coordinates"`.
+- Video and audio durations must not exceed 600 seconds (10 minutes).
+- The `speaker_frame`, `speaker_x`, and `speaker_y` parameters are only used when `speaker_selection` is set to "coordinates".
 
 ## Outputs
 
diff --git a/built-in-nodes/TextEncodeBooguEdit.mdx b/built-in-nodes/TextEncodeBooguEdit.mdx
index afc33e64b..5cdc22d1b 100644
--- a/built-in-nodes/TextEncodeBooguEdit.mdx
+++ b/built-in-nodes/TextEncodeBooguEdit.mdx
@@ -6,18 +6,20 @@ icon: "circle"
 mode: wide
 ---
 
-This node prepares conditioning for image editing with Boogu. It processes reference images to create both positive and negative conditioning outputs. The reference image is used twice: vision tokens from the image are added only to the positive conditioning to amplify the edit instruction, while a VAE reference latent is added to both positive and negative conditioning so it cancels out under CFG, preserving the original image identity. The tokenizer automatically selects the appropriate system prompt based on the presence of images and empty negative prompts.
+This node prepares conditioning for image editing with Boogu. It processes reference images to create both positive and negative conditioning outputs. The reference image is used twice: vision tokens from the image are added only to the positive conditioning to amplify the edit instruction, while a VAE reference latent is added to both positive and negative conditioning so it cancels out under CFG, preserving the original image identity; the tokenizer automatically selects the appropriate system prompt based on the presence of images and empty negative prompts.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `clip` | The CLIP model used for text encoding | CLIP | Yes | |
-| `prompt` | The text prompt describing the desired edit | STRING | Yes | |
-| `negative_prompt` | The text prompt describing what to avoid in the edit (can be empty to drop negative conditioning) | STRING | Yes | |
-| `vae` | The VAE model used to encode reference images into latent space | VAE | No | |
+| `prompt` | The text prompt describing the desired edit. Supports multiline text and dynamic prompts. | STRING | Yes | |
+| `negative_prompt` | The text prompt describing what to avoid in the edit. Can be left empty to drop negative conditioning. Advanced parameter. | STRING | Yes | |
+| `vae` | The VAE model used to encode reference images into latent space. Needed to add reference latents to the conditioning outputs. | VAE | No | |
 | `images` | Reference image(s) to edit. Boogu focuses on one reference per sample; more are allowed. | IMAGE | No | Up to 16 images |
 
+Reference latents are added to both conditioning outputs only when `vae` is provided together with at least one reference `image`. If `vae` is omitted, the positive output still receives vision tokens from the reference images, but neither output includes reference latents.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/TextEncodeJoyImageEdit.mdx b/built-in-nodes/TextEncodeJoyImageEdit.mdx
index 2a8a35e3c..b8db72a50 100644
--- a/built-in-nodes/TextEncodeJoyImageEdit.mdx
+++ b/built-in-nodes/TextEncodeJoyImageEdit.mdx
@@ -14,8 +14,10 @@ This node encodes a text prompt and optional images into conditioning data for u
 |-----------|-------------|-----------|----------|-------|
 | `clip` | The CLIP model used for encoding the text prompt | CLIP | Yes | - |
 | `prompt` | The text prompt to encode, supporting multiline input and dynamic prompts | STRING | Yes | - |
-| `vae` | A VAE model for encoding images into latent space (optional) | VAE | No | - |
-| `images` | One or more images to include in the conditioning, up to a maximum of 6 images | IMAGE | No | 0 to 6 images |
+| `vae` | A VAE model for encoding images into latent space | VAE | No | - |
+| `images` | Optional images to include in the conditioning, up to a maximum of 6 images. Each image connects as a separate input (image_0 to image_5) | IMAGE | No | 0 to 6 images |
+
+The `images` input is an autogrow group that accepts between 0 and 6 images. When no images are connected, the conditioning is based on the text prompt alone.
 
 ## Outputs
 
diff --git a/built-in-nodes/TextEncodeMageFlowEdit.mdx b/built-in-nodes/TextEncodeMageFlowEdit.mdx
index 33c908418..f4dca1ecb 100644
--- a/built-in-nodes/TextEncodeMageFlowEdit.mdx
+++ b/built-in-nodes/TextEncodeMageFlowEdit.mdx
@@ -16,17 +16,19 @@ This node encodes an edit instruction (prompt) along with one or more reference
 |-----------|-------------|-----------|----------|-------|
 | `clip` | The CLIP model used to tokenize and encode the text prompts. | CLIP | Yes | |
 | `prompt` | The editing instruction (positive prompt) to apply. | STRING | Yes | multiline, dynamic prompts enabled |
-| `negative_prompt` | The negative prompt to steer away from. Default: empty string (uses a space internally when blank). | STRING | No | multiline, dynamic prompts enabled |
+| `negative_prompt` | The negative prompt to steer away from. Default: empty string (uses a space internally when blank). Shown in the advanced section of the UI. | STRING | No | multiline, dynamic prompts enabled |
 | `vae` | VAE model to encode reference images into latent space. If not provided, no reference latents are added to the conditioning. | VAE | No | |
-| `images` | One or more reference images to edit. All images are resized to the output resolution before encoding. | IMAGE (autogrow) | No | Up to 16 images (named `image_1`…`image_16`), at least 0 |
+| `images` | Reference image(s) to edit. All references are resized to the output resolution before encoding. | IMAGE (autogrow) | No | Up to 16 images (named `image_1`…`image_16`), at least 0 |
 | `width` | Output width in pixels. If set to 0, the width of the first reference image is used. Always rounded down to a multiple of 16. Default: 0. | INT | Yes | 0 to 8192 (step 16) |
 | `height` | Output height in pixels. Same fallback behavior as width. Default: 0. | INT | Yes | 0 to 8192 (step 16) |
 | `batch_size` | Number of latent samples to generate. Default: 1. | INT | Yes | 1 to 4096 |
 
 **Notes on parameter dependencies:**
 - If `width` and/or `height` are 0 and no reference images are provided, they fall back to 1024 each.
+- If only one of `width` or `height` is 0, the missing dimension is taken from the first reference image while the explicitly set dimension is kept.
 - The `vae` parameter is optional; reference latents are only generated and attached to conditioning when a VAE is connected.
 - The `negative_prompt` field is optional – if left empty, a single space is used internally as the negative text.
+- For the text conditioning, each reference image is resized so its longest edge is at most 384 pixels, matching training preprocessing. The VAE encoding branch instead resizes all references to the full output resolution.
 
 ## Outputs
 
diff --git a/built-in-nodes/TextOverlay.mdx b/built-in-nodes/TextOverlay.mdx
index 1856f1edd..9e3eea024 100644
--- a/built-in-nodes/TextOverlay.mdx
+++ b/built-in-nodes/TextOverlay.mdx
@@ -13,13 +13,15 @@ This node draws text on top of an image or a batch of images. It creates a text
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `images` | The input image or batch of images to draw text on | IMAGE | Yes | |
-| `text` | The text to overlay on the image (default: "") | STRING | Yes | |
-| `font_size` | Font size as a percentage of the image height (default: 5.0) | FLOAT | Yes | 0.5 to 50.0 |
+| `text` | The text to overlay on the image (default: ""). Supports multiple lines; the `\n` and `\t` escape sequences are converted to newlines and tabs, and the text is automatically wrapped to fit within the image width. | STRING | Yes | |
+| `font_size` | Font size as a percentage of the image height (default: 5.0) | FLOAT | Yes | 0.5 to 50.0 (step 0.5) |
 | `color` | Color of the text (default: "#ffffff") | STRING | Yes | |
-| `position` | Vertical position of the text on the image (default: "top") | COMBO | Yes | `"top"`<br />`"bottom"` |
-| `align` | Horizontal alignment of the text (default: "left") | COMBO | Yes | `"left"`<br />`"center"`<br />`"right"` |
+| `position` | Vertical position of the text on the image (default: "top") | COMBO | Yes | "top"<br />"bottom" |
+| `align` | Horizontal alignment of the text (default: "left") | COMBO | Yes | "left"<br />"center"<br />"right" |
 | `outline` | Draw a black outline around the text (default: True) | BOOLEAN | Yes | |
 
+Note: If `text` is empty or contains only whitespace, the node returns the input images unchanged. The same text overlay is applied to every image in the batch.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/TripoImportModelNode.mdx b/built-in-nodes/TripoImportModelNode.mdx
index d3535c42f..53a9d4ee4 100644
--- a/built-in-nodes/TripoImportModelNode.mdx
+++ b/built-in-nodes/TripoImportModelNode.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-This node imports an external 3D model file into Tripo's system so you can use it with Tripo's post-processing nodes like Texture, Rig, and Convert. It uploads your model and returns a task ID that other Tripo nodes can use to reference the imported model. GLB format is recommended because textures are preserved only when embedded directly in the file. Note that texturing an imported model requires a texture prompt.
+This node imports an external 3D model into Tripo so that other Tripo post-processing nodes, such as Texture, Rig, and Convert, can use it. The node uploads the model and returns a task ID that identifies the imported model. GLB is recommended because textures are preserved only when embedded in the file, and texturing an imported model requires a texture prompt.
 
 ## Inputs
 
@@ -14,7 +14,7 @@ This node imports an external 3D model file into Tripo's system so you can use i
 |-----------|-------------|-----------|----------|-------|
 | `model_3d` | 3D model to import (GLB / FBX / OBJ / STL, up to 150 MB). OBJ and STL files carry no embedded textures. | FILE3D | Yes | GLB<br />FBX<br />OBJ<br />STL<br />Any 3D format |
 
-**Note:** GLB format is recommended because textures are preserved only when embedded directly in the file. OBJ and STL files do not support embedded textures. GLTF (.gltf) format is not supported because it references external files; use a single-file GLB instead. The file size must not exceed 150 MB.
+**Note:** Only GLB, FBX, OBJ, and STL formats are supported. GLTF (.gltf) cannot be imported because it references external files; use a single-file GLB instead. The model file must be 150 MB or smaller. GLB is recommended because textures survive import only when embedded in the file. OBJ and STL files do not carry embedded textures. Texturing an imported model requires a texture prompt.
 
 ## Outputs
 
diff --git a/built-in-nodes/VideoFrameSample.mdx b/built-in-nodes/VideoFrameSample.mdx
index 508b43f84..f2ac0a004 100644
--- a/built-in-nodes/VideoFrameSample.mdx
+++ b/built-in-nodes/VideoFrameSample.mdx
@@ -6,7 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
-The `VideoFrameSample` node extracts a fixed number of frames from a video using one of four strategies. For the contiguous strategies "head" and "tail", the output is a lazy video reference (frames are not decoded); for the non‑contiguous strategies "uniform" and "random", only the selected frames are decoded.
+The `VideoFrameSample` node extracts a fixed number of frames from a video using one of four strategies. For the contiguous strategies "head" and "tail", the output is a lazy video reference (frames are not decoded); for the non-contiguous strategies "uniform" and "random", only the selected frames are decoded.
 
 ## Inputs
 
@@ -14,11 +14,13 @@ The `VideoFrameSample` node extracts a fixed number of frames from a video using
 |-----------|-------------|-----------|----------|-------|
 | `video` | Input video. | VIDEO | Yes | – |
 | `num_frames` | Number of frames to sample (default: 16). | INT | Yes | 1 – 9999 |
-| `strategy` | Sampling strategy (default: "uniform"). | COMBO | Yes | `"uniform"`<br />`"head"`<br />`"tail"`<br />`"random"` |
+| `strategy` | uniform: evenly spaced, head: first N, tail: last N, random: random sorted (default: "uniform"). | COMBO | Yes | "uniform"<br />"head"<br />"tail"<br />"random" |
 | `seed` | Random seed, only used with the "random" strategy (default: 0). | INT | Yes | 0 – 18446744073709551615 |
 
 - `num_frames` is automatically clamped to the total frame count of the input video.
 - The `seed` parameter has no effect unless `strategy` is set to `"random"`.
+- When `strategy` is `"uniform"` and `num_frames` is 1, the middle frame of the video is selected.
+- An unrecognized `strategy` value raises an error.
 
 ## Outputs
 
diff --git a/built-in-nodes/VideoRandomTemporalCrop.mdx b/built-in-nodes/VideoRandomTemporalCrop.mdx
index b6e898617..f94ea20ca 100644
--- a/built-in-nodes/VideoRandomTemporalCrop.mdx
+++ b/built-in-nodes/VideoRandomTemporalCrop.mdx
@@ -6,14 +6,14 @@ icon: "circle"
 mode: wide
 ---
 
-Randomly crop a continuous range of frames from an input video. The crop length is controlled by the `length` parameter, and the start position is chosen using a random seed. The node operates lazily, meaning it does not process the entire video until the output is used downstream.
+Randomly crop a continuous range of frames from an input video. The number of frames to keep is set by the `length` parameter, and the start position is chosen randomly using the `seed` parameter. The node operates lazily, meaning it does not process the entire video until the output is used downstream.
 
 ## Inputs
 
 | Parameter | Description | Data Type | Required | Range |
 |-----------|-------------|-----------|----------|-------|
 | `video` | Input video. | VIDEO | Yes | – |
-| `length` | Number of frames to keep. (default: 16) | INT | Yes | min: 1, max: 99999 |
+| `length` | Number of frames to keep. If `length` is greater than the total number of frames in the video, the entire video is kept. (default: 16) | INT | Yes | min: 1, max: 99999 |
 | `seed` | Random seed. (default: 0) | INT | Yes | min: 0, max: 0xFFFFFFFFFFFFFFFF |
 
 ## Outputs
diff --git a/built-in-nodes/VideoTemporalCrop.mdx b/built-in-nodes/VideoTemporalCrop.mdx
index f595b8eac..9706ad773 100644
--- a/built-in-nodes/VideoTemporalCrop.mdx
+++ b/built-in-nodes/VideoTemporalCrop.mdx
@@ -16,6 +16,8 @@ This node crops a continuous range of frames from a video. It works in a fully l
 | `start_frame` | Starting frame index (default: 0). | INT | Yes | 0 to 99999 |
 | `length` | Number of frames to keep (default: 16). | INT | Yes | 1 to 99999 |
 
+Note: `start_frame` is limited to the last frame of the video, and `length` is reduced if it would extend beyond the available frames.
+
 ## Outputs
 
 | Output Name | Description | Data Type |
diff --git a/built-in-nodes/WanUni3CControlnetApply.mdx b/built-in-nodes/WanUni3CControlnetApply.mdx
index 38f2856dc..1e352b1cd 100644
--- a/built-in-nodes/WanUni3CControlnetApply.mdx
+++ b/built-in-nodes/WanUni3CControlnetApply.mdx
@@ -8,7 +8,7 @@ mode: wide
 
 ## Overview
 
-This node applies a Uni3C ControlNet to a Wan video diffusion model, using a rendered guidance video (e.g., warped point cloud renders) to influence the model’s output. It injects control signals at specific block layers, enabling camera‑trajectory‑based guidance during video generation.
+This node applies a Uni3C ControlNet to a Wan video diffusion model, using a rendered guidance video (e.g., warped point cloud renders) to influence the model's output. It injects control signals at specific block layers, enabling camera-trajectory-based guidance during video generation.
 
 ## Inputs
 
@@ -24,8 +24,10 @@ This node applies a Uni3C ControlNet to a Wan video diffusion model, using a ren
 
 **Notes:**
 - The `model_patch` must be a Uni3C ControlNet; otherwise the node raises an error.
-- The controlnet’s internal dimension must match the Wan model’s dimension – an error is raised if they differ.
+- The Uni3C ControlNet only works with Wan models; an error is raised if the model is not Wan-based.
+- The controlnet's internal dimension must match the Wan model's dimension – an error is raised if they differ.
 - The `render_video` input image is expected to be in RGB format (only the first 3 channels are used).
+- This node is experimental.
 
 ## Outputs
 
diff --git a/docs.json b/docs.json
index 1ffd2f6c5..f4ac0067b 100644
--- a/docs.json
+++ b/docs.json
@@ -2122,6 +2122,7 @@
                           {
                             "group": "Ideogram",
                             "pages": [
+                              "built-in-nodes/IdeogramPImage",
                               "built-in-nodes/IdeogramV3",
                               "built-in-nodes/IdeogramV4"
                             ]
@@ -2388,6 +2389,9 @@
                           {
                             "group": "Minimax",
                             "pages": [
+                              "built-in-nodes/MinimaxHailuo03FirstLastFrameNode",
+                              "built-in-nodes/MinimaxHailuo03ReferenceNode",
+                              "built-in-nodes/MinimaxHailuo03TextToVideoNode",
                               "built-in-nodes/MinimaxHailuoVideoNode",
                               "built-in-nodes/MinimaxImageToVideoNode",
                               "built-in-nodes/MinimaxSubjectToVideoNode",
@@ -4596,6 +4600,7 @@
                           {
                             "group": "Ideogram",
                             "pages": [
+                              "zh/built-in-nodes/IdeogramPImage",
                               "zh/built-in-nodes/IdeogramV3",
                               "zh/built-in-nodes/IdeogramV4"
                             ]
@@ -4862,6 +4867,9 @@
                           {
                             "group": "Minimax",
                             "pages": [
+                              "zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode",
+                              "zh/built-in-nodes/MinimaxHailuo03ReferenceNode",
+                              "zh/built-in-nodes/MinimaxHailuo03TextToVideoNode",
                               "zh/built-in-nodes/MinimaxHailuoVideoNode",
                               "zh/built-in-nodes/MinimaxImageToVideoNode",
                               "zh/built-in-nodes/MinimaxSubjectToVideoNode",
@@ -7436,6 +7444,7 @@
                           {
                             "group": "Ideogram",
                             "pages": [
+                              "ja/built-in-nodes/IdeogramPImage",
                               "ja/built-in-nodes/IdeogramV3",
                               "ja/built-in-nodes/IdeogramV4"
                             ]
@@ -7702,6 +7711,9 @@
                           {
                             "group": "Minimax",
                             "pages": [
+                              "ja/built-in-nodes/MinimaxHailuo03FirstLastFrameNode",
+                              "ja/built-in-nodes/MinimaxHailuo03ReferenceNode",
+                              "ja/built-in-nodes/MinimaxHailuo03TextToVideoNode",
                               "ja/built-in-nodes/MinimaxHailuoVideoNode",
                               "ja/built-in-nodes/MinimaxImageToVideoNode",
                               "ja/built-in-nodes/MinimaxSubjectToVideoNode",
@@ -10232,6 +10244,7 @@
                           {
                             "group": "Ideogram",
                             "pages": [
+                              "ko/built-in-nodes/IdeogramPImage",
                               "ko/built-in-nodes/IdeogramV3",
                               "ko/built-in-nodes/IdeogramV4"
                             ]
@@ -10498,6 +10511,9 @@
                           {
                             "group": "Minimax",
                             "pages": [
+                              "ko/built-in-nodes/MinimaxHailuo03FirstLastFrameNode",
+                              "ko/built-in-nodes/MinimaxHailuo03ReferenceNode",
+                              "ko/built-in-nodes/MinimaxHailuo03TextToVideoNode",
                               "ko/built-in-nodes/MinimaxHailuoVideoNode",
                               "ko/built-in-nodes/MinimaxImageToVideoNode",
                               "ko/built-in-nodes/MinimaxSubjectToVideoNode",
@@ -11814,4 +11830,4 @@
       "destination": "/ko/tutorials/partner-nodes/overview"
     }
   ]
-}
+}
\ No newline at end of file
diff --git a/ja/built-in-nodes/GeminiVideoOmni.mdx b/ja/built-in-nodes/GeminiVideoOmni.mdx
index 668728dab..d3a25a773 100644
--- a/ja/built-in-nodes/GeminiVideoOmni.mdx
+++ b/ja/built-in-nodes/GeminiVideoOmni.mdx
@@ -6,30 +6,32 @@ icon: "circle"
 mode: wide
 ---
 
-以下は、ご依頼いただいたComfyUIノードドキュメントの日本語訳です。
-
-GoogleのGemini Omni Flashモデルを使用して、テキストプロンプトから音声付きの動画を生成します。オプションで参照画像や動画を提供し、結果をガイドまたは編集することもできます。希望する長さ（3～10秒）とアスペクト比（16：9または9：16）をプロンプトに直接記述してください。
+Gemini Omni Flash モデルを使用して、テキストプロンプトから音声付きビデオを生成します。必要に応じて、参照画像や参照ビデオを提供して、結果をガイドまたは編集できます。希望する長さ（3〜10秒）とアスペクト比（16:9 または 9:16）をプロンプトに直接記述してください。
 
 ## 入力
 
 | パラメータ | 説明 | データ型 | 必須 | 範囲 |
 |-----------|-------------|-----------|----------|-------|
-| `model` | 動画生成に使用するGemini動画モデルです。 | MODEL | はい | "Omni Flash" |
-| `seed` | シードはノードを再実行するかどうかを制御します。シードに関わらず結果は非決定的です（デフォルト：42）。 | INT | はい | 0 ～ 2147483647 |
-| `prompt` | 生成する動画を説明するテキストプロンプトです。 先頭と末尾の空白を除去した後、少なくとも1文字以上の空白以外の文字が必要です。 | STRING | はい | 空白除去後、最低1文字 |
-| `images` | 動画生成をガイドするためのオプションの参照画像です。合計で最大14枚まで指定できます。 | IMAGE | いいえ | 複数画像可（最大14枚） |
-| `videos` | 動画生成をガイドまたは編集するためのオプションの参照動画です。各動画は最大10秒で、最大3本まで指定できます。 | VIDEO | いいえ | 複数動画可（最大3本、各最大10秒） |
-| `temperature` | 生成におけるランダム性を制御します（デフォルト：1.0）。 | FLOAT | いいえ | 0.0 ～ 2.0 |
-| `top_p` | 核サンプリングのパラメータです（デフォルト：0.95）。 | FLOAT | いいえ | 0.0 ～ 1.0 |
+| `model` | ビデオ生成に使用する Gemini ビデオモデル。 | COMBO | はい | "Omni Flash" |
+| `seed` | シードは、ノードを再実行するかどうかを制御します。結果はシードに関係なく非決定的です（デフォルト: 42）。 | INT | はい | 0 ～ 2147483647 |
+| `prompt` | 生成するビデオを説明するテキストプロンプト。先頭と末尾の空白を除去した後、少なくとも1文字以上の空白以外の文字が必要です。 | STRING | はい | 空白除去後、最低1文字 |
+| `images` | ビデオ生成をガイドするためのオプションの参照画像。合計で最大14枚です。 | IMAGE | いいえ | 複数画像可（最大14枚） |
+| `videos` | ビデオ生成をガイドまたは編集するためのオプションの参照ビデオ。各最大10秒、最大3本です。 | VIDEO | いいえ | 複数ビデオ可（最大3本、各最大10秒） |
+| `temperature` | 生成におけるランダム性を制御します（デフォルト: 1.0）。 | FLOAT | いいえ | 0.0 ～ 2.0 |
+| `top_p` | 核サンプリングのパラメータ（デフォルト: 0.95）。 | FLOAT | いいえ | 0.0 ～ 1.0 |
+
+注記：
+- 画像入力に複数のフレームが含まれる場合、各フレームが最大14枚の画像にカウントされます。
+- `images` または `videos` が指定された場合、結合されたエンコード済みメディアのサイズは約90MB未満に保つ必要があります。それ以外の場合、ノードはエラーを発生させます。
 
 ## 出力
 
 | 出力名 | 説明 | データ型 |
 |-------------|-------------|-----------|
-| `VIDEO` | Geminiモデルから生成された音声付き動画です。 | VIDEO |
-| `STRING` | 推論や説明など、モデルからのテキスト応答です。 | STRING |
+| `VIDEO` | Gemini モデルから生成された音声付きビデオ。 | VIDEO |
+| `STRING` | 推論や説明など、モデルからのテキスト応答。 | STRING |
 
 > このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください！ [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/GeminiVideoOmni/ja.md)
 
 ---
-**Source fingerprint (SHA-256):** `046842b7ec736283bba355aaa038b02fcf2416020f5f7aee7b0150d2a05bcbe6`
+**Source fingerprint (SHA-256):** `1b7ca51d07cfb6a166cfed2a7e7174fd62f3290abcc1bdfdce94369dda242d3f`
diff --git a/ja/built-in-nodes/IdeogramPImage.mdx b/ja/built-in-nodes/IdeogramPImage.mdx
new file mode 100644
index 000000000..591d6ecf6
--- /dev/null
+++ b/ja/built-in-nodes/IdeogramPImage.mdx
@@ -0,0 +1,34 @@
+---
+title: "IdeogramPImage - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the IdeogramPImage node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "IdeogramPImage"
+icon: "circle"
+mode: wide
+---
+
+Ideogram P-Image は、Ideogram の高速テキスト画像変換モデルを使用して、テキストプロンプトから画像を生成します。このモデルは、優れたタイポグラフィとフォトリアリズムで知られています。また、Ideogram 4.0 の構造化 JSON キャプションにも対応しており、テキスト文字列、色、レイアウトを正確に制御できます。このノードは、生成された画像と、画像が実際に生成された元の最終プロンプトを返します。
+
+## 入力
+
+| パラメータ | 説明 | データ型 | 必須 | 範囲 |
+|-----------|-------------|-----------|----------|-------|
+| `prompt` | テキストプロンプト。Ideogram 4.0 の構造化 JSON キャプション（#RRGGBB 形式の正確な色、正確なテキスト文字列、バウンディングボックスレイアウト）も受け入れます。そのまま使用するには、`prompt_upsampling` を OFF に設定してください。空にすることはできません。（デフォルト: ""） | STRING | はい | 任意のテキスト |
+| `quality` | 速度/コスト/品質の階層です。MEDIUM は日常的なデフォルトです。HIGH は複雑なプロンプト、微細なディテール、難しいテキスト向けです。VERY_LOW/LOW は大規模な下書き用です。難しいテキストは MEDIUM 未満では描画品質が低下します。（デフォルト: "MEDIUM"） | STRING | はい | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
+| `resolution` | 出力サイズのクラスです（正確なピクセル数はアスペクト比に応じて決まります。例: 16:9 の場合、1K では 1280x720、2K では 2560x1440 になります）。鮮明なタイポグラフィが必要な場合は、HIGH + 2K を推奨します。（デフォルト: "1K"） | STRING | はい | "1K"<br />"2K" |
+| `aspect_ratio` | 画像生成時のアスペクト比です。（デフォルト: "1:1"） | STRING | はい | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
+| `prompt_upsampling` | 生成前に短いプロンプトを詳細な構造化キャプションに展開します（書き換えられたプロンプトは `final_prompt` として返されます）。独自の JSON キャプションや正確な文言を指定する場合は OFF に設定してください。（デフォルト: "AUTO"） | STRING | はい | "AUTO"<br />"ON"<br />"OFF" |
+| `seed` | 再現可能な生成のためのシードです。`prompt_upsampling` が OFF の場合、同じシードと設定で同じ画像が返されます。ON/AUTO の場合、プロンプトの書き換えは実行ごとに変化します。結果を再現するには、その `final_prompt` 出力を `prompt_upsampling` OFF かつ同じシードで再利用してください。（デフォルト: 42） | INT | いいえ | 0 から 2147483647 |
+
+**制約に関する注意:** プロンプトには空白以外の文字が少なくとも 1 文字含まれている必要があります。含まれていない場合、ノードは失敗します。独自の構造化 JSON キャプションや正確な文言を指定する場合は、`prompt_upsampling` を OFF に設定してください。`prompt_upsampling` が ON または AUTO の場合、生成前にプロンプトが書き換えられるため、同じシードでも同じ画像が再現されないことがあります。画像を再現するには、その `final_prompt` 出力を `prompt_upsampling` OFF かつ同じシードで再利用してください。
+
+## 出力
+
+| 出力名 | 説明 | データ型 |
+|-------------|-------------|-----------|
+| `image` | 生成された画像です。画像のバッチとして返されます。Ideogram のコンテンツセーフティフィルターが生成をブロックした場合は、代わりにエラーが発生します。 | IMAGE |
+| `final_prompt` | 画像が実際に生成された元のプロンプトです（`prompt_upsampling` が実行された場合は書き換え後の構造化キャプション、それ以外の場合は指定したプロンプト）。この画像を再現するには、`prompt_upsampling` を OFF にし、同じシードでこのプロンプトを再度入力してください。 | STRING |
+
+> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください！ [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/IdeogramPImage/ja.md)
+
+---
+**Source fingerprint (SHA-256):** `7bd20aae508fee111ded32e87119ed6fc01c5ad5ba7d595e24391830a0f20bb7`
diff --git a/ja/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx b/ja/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
new file mode 100644
index 000000000..f83fd8039
--- /dev/null
+++ b/ja/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
@@ -0,0 +1,36 @@
+---
+title: "MinimaxHailuo03FirstLastFrameNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03FirstLastFrameNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03FirstLastFrameNode"
+icon: "circle"
+mode: wide
+---
+
+このノードは、最初のフレーム画像と、必要に応じて最後のフレーム画像を使用して、MiniMax H3 モデルで動画を生成します。動画は指定された画像のアスペクト比に従い、最後のフレームが指定された場合は、最初のフレームから最後のフレームに向かってアニメーションします。
+
+## 入力
+
+| パラメータ | 説明 | データ型 | 必須 | 範囲 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 動画生成に使用するモデルです。このコンボには、モデル選択（"MiniMax H3"）、生成する動画を説明するテキストプロンプト、出力解像度、動画の長さが含まれます。プロンプトには、少なくとも1文字以上の非空白文字が含まれている必要があります。 | COMBO | はい | "MiniMax H3" |
+| `first_frame` | 動画の最初のフレーム画像です。生成される動画のアスペクト比は、この画像に従います。少なくとも256x256ピクセル以上で、幅と高さのアスペクト比が0.4〜2.5の範囲内である必要があります。 | IMAGE | はい | - |
+| `last_frame` | 動画の最後のフレーム画像（オプション）です。指定すると、動画は最初のフレームから始まり、この画像で終わります。`first_frame` と同じサイズおよびアスペクト比の要件を満たす必要があります。 | IMAGE | いいえ | - |
+| `seed` | ランダムシードです。同じシードで同じリクエストを行うと、類似した結果が得られますが、完全に同一であるとは限りません。各生成後にランダム化する「control after generate」オプションが含まれています。デフォルト: 42。 | INT | はい | 0 から 4294967295 |
+| `watermark` | 動画に AIGC 透かしを追加するかどうかです。これは詳細設定パラメータです。デフォルト: False。 | BOOLEAN | はい | True<br />False |
+
+**制約事項:**
+- `model` コンボ内のテキストプロンプトは空にできません。空白のみのプロンプトは拒否されます。
+- 指定するフレーム画像（`first_frame` および、使用する場合は `last_frame`）は、幅 256 ピクセル以上、高さ 256 ピクセル以上で、幅と高さのアスペクト比が 0.4 〜 2.5（約 2:5 〜 5:2）の範囲内である必要があります。
+- `last_frame` はオプションです。省略した場合、動画は最初のフレームのみから生成されます。
+- 出力動画のアスペクト比は、指定された画像によって決定され、別途の比率設定はありません。
+
+## 出力
+
+| 出力名 | 説明 | データ型 |
+|-------------|-------------|-----------|
+| `video` | 最初のフレームとオプションの最後のフレームから、MiniMax H3 モデルを使用して生成された動画です。 | VIDEO |
+
+> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください！ [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03FirstLastFrameNode/ja.md)
+
+---
+**Source fingerprint (SHA-256):** `f4cb9217eb346019680c64b30c1beacce16f0050616b7b76265edc5840f6b21e`
diff --git a/ja/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx b/ja/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
new file mode 100644
index 000000000..c3301bfe9
--- /dev/null
+++ b/ja/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
@@ -0,0 +1,42 @@
+---
+title: "MinimaxHailuo03ReferenceNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03ReferenceNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03ReferenceNode"
+icon: "circle"
+mode: wide
+---
+
+このノードは、MiniMax H3 モデルを使用して動画を生成します。参照画像、動画、音声を条件として結果を制御します。参照はプロンプト内で接続順に「Image 1」「Image 2」「Video 1」「Audio 1」のように指定されます。
+
+## 入力
+
+| パラメータ | 説明 | データ型 | 必須 | 範囲 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 動画生成に使用するモデル（デフォルト: "MiniMax H3"）。"MiniMax H3" を選択すると、以下の `prompt`、`duration`、`resolution`、`ratio`、`reference_images`、`reference_videos`、`reference_audios` の設定が提供されます。 | STRING | はい | "MiniMax H3" |
+| `prompt` | 生成する動画のテキストによる説明。参照メディアは順序で指定できます。例: "Image 1"、"Image 2"、"Video 1"、"Audio 1"。 | STRING | はい | 最小長: 1文字 |
+| `duration` | 生成する動画の長さ（秒）。 | INT | はい | 複数のオプションから選択可能 |
+| `resolution` | 生成する動画の出力解像度。 | STRING | はい | 複数のオプションから選択可能 |
+| `ratio` | 生成する動画のアスペクト比。 | STRING | はい | 複数のオプションから選択可能 |
+| `reference_images` | 被写体またはスタイルの参照画像。プロンプト内で接続順に "Image 1" ～ "Image 9" として指定されます。最大9枚。 | IMAGE | いいえ | 0〜9枚 |
+| `reference_videos` | 動きまたはシーンの参照動画。プロンプト内で接続順に "Video 1" ～ "Video 3" として指定されます。最大3本、各2〜15秒、合計15秒以内。 | VIDEO | いいえ | 0〜3本 |
+| `reference_audios` | 音声参照。プロンプト内で接続順に "Audio 1" ～ "Audio 3" として指定されます。最大3クリップ、各2〜15秒、合計15秒以内。参照画像または参照動画なしでは使用できません。 | AUDIO | いいえ | 0〜3クリップ |
+| `seed` | 乱数シード。同じシードで同じリクエストを行うと類似した結果が得られますが、完全に同一の結果は保証されません（デフォルト: 42）。 | INT | はい | 0〜4294967295 |
+| `watermark` | 動画にAIGC透かしを追加するかどうか（デフォルト: false）。 | BOOLEAN | いいえ | true<br />false |
+
+### パラメータの制約
+
+- 参照画像または参照動画が少なくとも1つ必要です。参照音声のみでは受け付けられません。
+- 各参照画像のアスペクト比は約0.4〜2.5（2:5〜5:2）の範囲内で、幅と高さがそれぞれ最低256ピクセルである必要があります。
+- 各参照動画は2〜15秒の長さで、フレームレートが23.976〜60 FPSの範囲内である必要があります。すべての参照動画の合計時間は15秒を超えることはできません。
+- 各参照音声クリップは2〜15秒の長さである必要があります。すべての参照音声クリップの合計時間は15秒を超えることはできません。
+
+## 出力
+
+| 出力名 | 説明 | データ型 |
+|-------------|-------------|-----------|
+| `video` | 生成された動画。 | VIDEO |
+
+> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください！ [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03ReferenceNode/ja.md)
+
+---
+**Source fingerprint (SHA-256):** `beca020333a544188e6c21829eb8e63415aa5299efc676438e85662a5f08660d`
diff --git a/ja/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx b/ja/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
new file mode 100644
index 000000000..e03f072b9
--- /dev/null
+++ b/ja/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
@@ -0,0 +1,30 @@
+---
+title: "MinimaxHailuo03TextToVideoNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03TextToVideoNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03TextToVideoNode"
+icon: "circle"
+mode: wide
+---
+
+このノードは、MiniMax H3モデルを使用してテキストプロンプトから動画を生成します。テキストを解像度、再生時間、アスペクト比などの動画設定とともにMiniMax APIに送信し、生成された動画を出力として返します。
+
+## 入力
+
+| パラメータ | 説明 | データ型 | 必須 | 範囲 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 動画生成に使用するモデル。（デフォルト："MiniMax H3"）この選択には、生成される動画のテキストプロンプト、解像度、再生時間、アスペクト比の設定も含まれます。 | COMBO | はい | `"MiniMax H3"` |
+| `seed` | 乱数シード。同じシードで同じリクエストを送信すると、類似した結果が得られますが、同一の結果が保証されるわけではありません。（デフォルト：42） | INT | はい | 0〜4294967295 |
+| `watermark` | 動画にAIGCウォーターマークを追加するかどうか。（デフォルト：false） | BOOLEAN | いいえ | true<br />false |
+
+注：`model` オプションに含まれるテキストプロンプトには、少なくとも1つの空白以外の文字が含まれている必要があります。このノードに表示される推定価格は、選択された動画の再生時間から計算されます。
+
+## 出力
+
+| 出力名 | 説明 | データ型 |
+|-------------|-------------|-----------|
+| `VIDEO` | 提供されたテキストプロンプトから生成された動画。 | VIDEO |
+
+> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください！ [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03TextToVideoNode/ja.md)
+
+---
+**Source fingerprint (SHA-256):** `9478576dd02ed407a39c95c7227eb8e1482db8b77adc814691fbd807e4cc2893`
diff --git a/ko/built-in-nodes/GeminiVideoOmni.mdx b/ko/built-in-nodes/GeminiVideoOmni.mdx
index 49b0f964d..a090fed65 100644
--- a/ko/built-in-nodes/GeminiVideoOmni.mdx
+++ b/ko/built-in-nodes/GeminiVideoOmni.mdx
@@ -6,28 +6,32 @@ icon: "circle"
 mode: wide
 ---
 
-Google의 Gemini Omni Flash 모델을 사용하여 텍스트 프롬프트로 오디오가 포함된 비디오를 생성합니다. 필요에 따라 참조 이미지 및/또는 비디오를 제공하여 결과를 안내하거나 편집할 수 있습니다. 원하는 길이(3~10초)와 화면 비율(16:9 또는 9:16)을 프롬프트에 직접 설명하세요.
+Google의 Gemini Omni Flash 모델을 사용하여 텍스트 프롬프트에서 오디오가 포함된 비디오를 생성합니다. 선택적으로 참조 이미지 및/또는 비디오를 제공하여 결과를 안내하거나 편집할 수 있습니다. 원하는 길이(3-10초)와 화면 비율(16:9 또는 9:16)을 프롬프트에 직접 설명하세요.
 
 ## 입력
 
-| 매개변수 | 설명 | 데이터 타입 | 필수 여부 | 범위 |
+| 매개변수 | 설명 | 데이터 타입 | 필수 | 범위 |
 |-----------|-------------|-----------|----------|-------|
-| `model` | 비디오 생성에 사용되는 Gemini 비디오 모델입니다. | MODEL | 예 | "Omni Flash" |
-| `seed` | 시드는 노드 재실행 여부를 제어합니다. 시드와 관계없이 결과는 비결정적입니다(기본값: 42). | INT | 예 | 0 ~ 2147483647 |
-| `prompt` | 생성할 비디오를 설명하는 텍스트 프롬프트입니다. 앞뒤 공백을 제거한 후 최소 하나 이상의 공백이 아닌 문자가 있어야 합니다. | STRING | 예 | 공백 제거 후 최소 1자 |
-| `images` | 비디오 생성을 안내하는 선택적 참조 이미지입니다. 최대 14개의 이미지를 사용할 수 있습니다. | IMAGE | 아니요 | 여러 이미지 허용(최대 14개) |
-| `videos` | 비디오 생성을 안내하거나 편집하는 선택적 참조 비디오입니다. 각각 최대 10초 길이의 비디오를 최대 3개까지 사용할 수 있습니다. | VIDEO | 아니요 | 여러 비디오 허용(최대 3개, 각 최대 10초) |
-| `temperature` | 생성 과정의 무작위성을 제어합니다(기본값: 1.0). | FLOAT | 아니요 | 0.0 ~ 2.0 |
-| `top_p` | 핵 샘플링 매개변수입니다(기본값: 0.95). | FLOAT | 아니요 | 0.0 ~ 1.0 |
+| `model` | 비디오 생성을 위해 사용되는 Gemini 비디오 모델입니다. | COMBO | 예 | "Omni Flash" |
+| `seed` | 시드는 노드가 다시 실행되어야 하는지 여부를 제어합니다. 시드와 관계없이 결과는 비결정적입니다(기본값: 42). | INT | 예 | 0 ~ 2147483647 |
+| `prompt` | 생성할 비디오를 설명하는 텍스트 프롬프트입니다. 앞뒤 공백을 제거한 후 공백이 아닌 문자가 하나 이상 있어야 합니다. | STRING | 예 | 공백 제거 후 최소 1자 |
+| `images` | 비디오 생성을 안내하기 위한 선택적 참조 이미지입니다. 총 최대 14개의 이미지를 사용할 수 있습니다. | IMAGE | 아니요 | 여러 이미지 허용(최대 14개) |
+| `videos` | 비디오 생성을 안내하거나 편집하기 위한 선택적 참조 비디오입니다. 최대 3개의 비디오를 사용할 수 있으며, 각 비디오는 최대 10초입니다. | VIDEO | 아니요 | 여러 비디오 허용(최대 3개, 각 최대 10초) |
+| `temperature` | 생성 시 무작위성을 제어합니다(기본값: 1.0). | FLOAT | 아니요 | 0.0 ~ 2.0 |
+| `top_p` | 핵심 샘플링 매개변수입니다(기본값: 0.95). | FLOAT | 아니요 | 0.0 ~ 1.0 |
+
+참고:
+- 이미지 입력에 여러 프레임이 포함된 경우 각 프레임이 최대 14개 이미지의 개수에 포함됩니다.
+- `images` 또는 `videos`가 제공되는 경우, 인코딩된 미디어의 총 크기가 약 90MB 미만이어야 합니다. 그렇지 않으면 노드에서 오류가 발생합니다.
 
 ## 출력
 
 | 출력 이름 | 설명 | 데이터 타입 |
 |-------------|-------------|-----------|
 | `VIDEO` | Gemini 모델에서 생성된 오디오가 포함된 비디오입니다. | VIDEO |
-| `STRING` | 추론 또는 설명과 같은 모델의 모든 텍스트 응답입니다. | STRING |
+| `STRING` | 추론이나 설명과 같은 모델의 텍스트 응답입니다. | STRING |
 
 > 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/GeminiVideoOmni/ko.md)
 
 ---
-**Source fingerprint (SHA-256):** `046842b7ec736283bba355aaa038b02fcf2416020f5f7aee7b0150d2a05bcbe6`
+**Source fingerprint (SHA-256):** `1b7ca51d07cfb6a166cfed2a7e7174fd62f3290abcc1bdfdce94369dda242d3f`
diff --git a/ko/built-in-nodes/IdeogramPImage.mdx b/ko/built-in-nodes/IdeogramPImage.mdx
new file mode 100644
index 000000000..d0dfed7ff
--- /dev/null
+++ b/ko/built-in-nodes/IdeogramPImage.mdx
@@ -0,0 +1,34 @@
+---
+title: "IdeogramPImage - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the IdeogramPImage node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "IdeogramPImage"
+icon: "circle"
+mode: wide
+---
+
+Ideogram P-Image는 Ideogram의 빠른 텍스트-이미지 모델을 사용하여 텍스트 프롬프트에서 이미지를 생성합니다. 이 모델은 뛰어난 타이포그래피와 사실적인 이미지 생성으로 잘 알려져 있습니다. 또한 Ideogram 4.0 구조화된 JSON 캡션을 지원하여 텍스트 문자열, 색상 및 레이아웃을 정확하게 제어할 수 있습니다. 이 노드는 생성된 이미지와 함께 이미지가 실제로 생성된 최종 프롬프트를 반환합니다.
+
+## 입력
+
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `prompt` | 텍스트 프롬프트. Ideogram 4.0 구조화된 JSON 캡션(정확한 색상은 #RRGGBB 16진수, 정확한 텍스트 문자열, 경계 상자 레이아웃)도 허용됩니다. 해당 캡션을 그대로 사용하려면 prompt_upsampling을 OFF로 설정하세요. 비어 있지 않아야 합니다. (기본값: "") | STRING | 예 | 모든 텍스트 |
+| `quality` | 속도/가격/품질 등급입니다. MEDIUM은 일상적인 기본값이며, HIGH는 복잡한 프롬프트, 미세한 디테일, 어려운 텍스트에 적합합니다. VERY_LOW/LOW는 대량 초안 제작용입니다. 어려운 텍스트는 MEDIUM 미만에서는 제대로 렌더링되지 않습니다. (기본값: "MEDIUM") | STRING | 예 | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
+| `resolution` | 출력 크기 등급입니다. (정확한 픽셀 수는 종횡비를 따릅니다. 예: 16:9는 1K에서 1280x720, 2K에서 2560x1440을 제공합니다.) 선명한 타이포그래피를 위해서는 HIGH + 2K를 권장합니다. (기본값: "1K") | STRING | 예 | "1K"<br />"2K" |
+| `aspect_ratio` | 이미지 생성의 종횡비입니다. (기본값: "1:1") | STRING | 예 | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
+| `prompt_upsampling` | 생성 전에 짧은 프롬프트를 상세한 구조화된 캡션으로 확장합니다. (재작성된 프롬프트는 final_prompt로 반환됩니다.) 자체 JSON 캡션이나 정확한 문구를 제공하는 경우에는 OFF로 설정하세요. (기본값: "AUTO") | STRING | 예 | "AUTO"<br />"ON"<br />"OFF" |
+| `seed` | 재현 가능한 생성을 위한 시드입니다. prompt_upsampling이 OFF이면 동일한 시드와 설정으로 동일한 이미지가 반환됩니다. ON/AUTO이면 프롬프트 재작성이 실행마다 달라지므로, 결과를 재현하려면 해당 final_prompt 출력을 prompt_upsampling OFF 및 동일한 시드와 함께 재사용하세요. (기본값: 42) | INT | 아니요 | 0 ~ 2147483647 |
+
+**제약 조건 참고 사항:** 프롬프트에는 공백이 아닌 문자가 하나 이상 포함되어야 합니다. 그렇지 않으면 노드가 실패합니다. 자체 구조화된 JSON 캡션이나 정확한 문구를 제공하는 경우 `prompt_upsampling`을 OFF로 설정하세요. `prompt_upsampling`이 ON 또는 AUTO인 경우 생성 전에 프롬프트가 재작성되므로 동일한 시드를 사용하더라도 동일한 이미지가 재현되지 않을 수 있습니다. 이미지를 재현하려면 해당 `final_prompt` 출력을 `prompt_upsampling` OFF 및 동일한 시드와 함께 재사용하세요.
+
+## 출력
+
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `image` | 생성된 이미지로, 이미지 배치 형태로 반환됩니다. Ideogram의 콘텐츠 안전 필터가 생성을 차단하면 대신 오류가 발생합니다. | IMAGE |
+| `final_prompt` | 이미지가 실제로 생성된 프롬프트입니다. (prompt_upsampling이 실행된 경우 재작성된 구조화된 캡션이고, 그렇지 않은 경우 사용자의 프롬프트입니다.) 이 값을 prompt_upsampling OFF 및 동일한 시드와 함께 다시 입력하면 이 이미지를 재현할 수 있습니다. | STRING |
+
+> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/IdeogramPImage/ko.md)
+
+---
+**Source fingerprint (SHA-256):** `7bd20aae508fee111ded32e87119ed6fc01c5ad5ba7d595e24391830a0f20bb7`
diff --git a/ko/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx b/ko/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
new file mode 100644
index 000000000..e6c250ab9
--- /dev/null
+++ b/ko/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
@@ -0,0 +1,36 @@
+---
+title: "MinimaxHailuo03FirstLastFrameNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03FirstLastFrameNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03FirstLastFrameNode"
+icon: "circle"
+mode: wide
+---
+
+이 노드는 MiniMax H3 모델을 사용하여 첫 번째 프레임 이미지와 선택적 마지막 프레임 이미지로부터 비디오를 생성합니다. 비디오는 제공된 이미지의 종횡비를 따르며, 마지막 프레임이 제공되면 첫 번째 프레임에서 마지막 프레임 방향으로 애니메이션됩니다.
+
+## 입력
+
+| 매개변수 | 설명 | 데이터 타입 | 필수 | 범위 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 비디오 생성에 사용할 모델입니다. 이 콤보에는 모델 선택("MiniMax H3"), 생성할 비디오를 설명하는 텍스트 프롬프트, 출력 해상도 및 비디오 길이가 포함됩니다. 프롬프트에는 공백이 아닌 문자가 하나 이상 포함되어야 합니다. | COMBO | 예 | "MiniMax H3" |
+| `first_frame` | 비디오의 첫 번째 프레임 이미지입니다. 생성된 비디오의 종횡비는 이 이미지를 따릅니다. 최소 256x256 픽셀이어야 하며 가로-세로 종횡비는 0.4에서 2.5 사이여야 합니다. | IMAGE | 예 | - |
+| `last_frame` | 비디오의 선택적 마지막 프레임 이미지입니다. 제공되면 비디오는 첫 번째 프레임에서 시작하여 이 이미지에서 끝납니다. `first_frame`과 동일한 크기 및 종횡비 요구 사항을 충족해야 합니다. | IMAGE | 아니요 | - |
+| `seed` | 난수 시드입니다. 동일한 시드를 사용한 동일한 요청은 유사한 결과를 제공하지만 동일한 결과를 보장하지는 않습니다. 각 생성 후 무작위화하는 "control after generate" 옵션이 포함됩니다. 기본값: 42. | INT | 예 | 0 ~ 4294967295 |
+| `watermark` | 비디오에 AIGC 워터마크를 추가할지 여부입니다. 이는 고급 매개변수입니다. 기본값: False. | BOOLEAN | 예 | True<br />False |
+
+**제약 사항 참고:**
+- `model` 콤보 내부의 텍스트 프롬프트는 비어 있을 수 없습니다. 공백만 있는 프롬프트는 거부됩니다.
+- 제공되는 모든 프레임 이미지(`first_frame` 및 사용 시 `last_frame`)는 최소 256x256 픽셀이어야 하며, 가로-세로 종횡비가 0.4에서 2.5 사이(약 2:5 ~ 5:2)여야 합니다.
+- `last_frame`은 선택 사항입니다. 생략하면 첫 번째 프레임에서만 비디오가 생성됩니다.
+- 출력 비디오의 종횡비는 별도의 비율 설정이 아닌 제공된 이미지에 의해 결정됩니다.
+
+## 출력
+
+| 출력 이름 | 설명 | 데이터 타입 |
+|-------------|-------------|-----------|
+| `video` | MiniMax H3 모델을 사용하여 첫 번째 프레임과 선택적 마지막 프레임으로 생성된 비디오입니다. | VIDEO |
+
+> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03FirstLastFrameNode/ko.md)
+
+---
+**Source fingerprint (SHA-256):** `f4cb9217eb346019680c64b30c1beacce16f0050616b7b76265edc5840f6b21e`
diff --git a/ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx b/ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
new file mode 100644
index 000000000..382969029
--- /dev/null
+++ b/ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
@@ -0,0 +1,42 @@
+---
+title: "MinimaxHailuo03ReferenceNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03ReferenceNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03ReferenceNode"
+icon: "circle"
+mode: wide
+---
+
+이 노드는 MiniMax H3 모델을 사용하여 비디오를 생성하며, 참조 이미지, 비디오 및 오디오를 사용하여 결과를 조건화합니다. 참조 항목은 프롬프트에서 연결 순서에 따라 "Image 1", "Image 2", "Video 1", "Audio 1" 등으로 지칭됩니다.
+
+## 입력
+
+| 매개변수 | 설명 | 데이터 타입 | 필수 여부 | 범위 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 비디오 생성에 사용할 모델입니다 (기본값: "MiniMax H3"). "MiniMax H3"를 선택하면 아래의 `prompt`, `duration`, `resolution`, `ratio`, `reference_images`, `reference_videos`, `reference_audios` 설정이 제공됩니다. | STRING | 예 | "MiniMax H3" |
+| `prompt` | 생성할 비디오에 대한 텍스트 설명입니다. 참조 미디어는 순서에 따라 예를 들어 "Image 1", "Image 2", "Video 1" 또는 "Audio 1"로 지칭할 수 있습니다. | STRING | 예 | 최소 길이: 1자 |
+| `duration` | 생성된 비디오의 길이(초)입니다. | INT | 예 | 여러 옵션 제공 |
+| `resolution` | 생성된 비디오의 출력 해상도입니다. | STRING | 예 | 여러 옵션 제공 |
+| `ratio` | 생성된 비디오의 화면 비율입니다. | STRING | 예 | 여러 옵션 제공 |
+| `reference_images` | 피사체 또는 스타일 참조 이미지로, 연결 순서에 따라 프롬프트에서 "Image 1".."Image 9"로 지칭됩니다. 최대 9개까지 사용할 수 있습니다. | IMAGE | 아니요 | 0~9개 이미지 |
+| `reference_videos` | 모션 또는 장면 참조 비디오로, 연결 순서에 따라 프롬프트에서 "Video 1".."Video 3"로 지칭됩니다. 최대 3개까지 사용할 수 있으며, 각각 2~15초, 총 15초입니다. | VIDEO | 아니요 | 0~3개 비디오 |
+| `reference_audios` | 오디오 참조 항목으로, 연결 순서에 따라 프롬프트에서 "Audio 1".."Audio 3"로 지칭됩니다. 최대 3개 클립, 각각 2~15초, 총 15초입니다. 참조 이미지 또는 비디오 없이 사용할 수 없습니다. | AUDIO | 아니요 | 0~3개 클립 |
+| `seed` | 난수 시드입니다. 동일한 시드를 사용한 동일한 요청은 유사하지만 완전히 동일하다고 보장되지 않는 결과를 제공합니다 (기본값: 42). | INT | 예 | 0 ~ 4294967295 |
+| `watermark` | 비디오에 AIGC 워터마크를 추가할지 여부입니다 (기본값: false). | BOOLEAN | 아니요 | true<br />false |
+
+### 매개변수 제약 조건
+
+- 최소한 하나의 참조 이미지 또는 하나의 참조 비디오가 필요합니다. 참조 오디오만으로는 사용할 수 없습니다.
+- 각 참조 이미지의 화면 비율은 약 0.4~2.5 (2:5 ~ 5:2) 사이여야 하며, 최소 너비와 높이는 256픽셀이어야 합니다.
+- 각 참조 비디오의 길이는 2~15초 사이여야 하며, 프레임 레이트는 23.976~60 FPS 사이여야 합니다. 모든 참조 비디오의 총 길이는 15초를 초과할 수 없습니다.
+- 각 참조 오디오 클립의 길이는 2~15초 사이여야 합니다. 모든 참조 오디오 클립의 총 길이는 15초를 초과할 수 없습니다.
+
+## 출력
+
+| 출력 이름 | 설명 | 데이터 타입 |
+|-------------|-------------|-----------|
+| `video` | 생성된 비디오입니다. | VIDEO |
+
+> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03ReferenceNode/ko.md)
+
+---
+**Source fingerprint (SHA-256):** `beca020333a544188e6c21829eb8e63415aa5299efc676438e85662a5f08660d`
diff --git a/ko/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx b/ko/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
new file mode 100644
index 000000000..84795de77
--- /dev/null
+++ b/ko/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
@@ -0,0 +1,30 @@
+---
+title: "MinimaxHailuo03TextToVideoNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03TextToVideoNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03TextToVideoNode"
+icon: "circle"
+mode: wide
+---
+
+이 노드는 MiniMax H3 모델을 사용하여 텍스트 프롬프트에서 비디오를 생성합니다. 텍스트와 해상도, 지속 시간, 화면 비율 등의 비디오 설정을 MiniMax API로 전송하고, 생성된 비디오를 출력으로 반환합니다.
+
+## 입력
+
+| 매개변수 | 설명 | 데이터 타입 | 필수 | 범위 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 비디오 생성에 사용할 모델입니다. (기본값: "MiniMax H3"). 이 선택 항목에는 생성된 비디오의 텍스트 프롬프트, 해상도, 지속 시간 및 화면 비율 설정도 포함됩니다. | COMBO | 예 | `"MiniMax H3"` |
+| `seed` | 난수 시드입니다. 동일한 시드를 사용한 동일한 요청은 유사한 결과를 제공하지만, 완전히 동일한 결과는 보장되지 않습니다. (기본값: 42) | INT | 예 | 0 ~ 4294967295 |
+| `watermark` | 비디오에 AIGC 워터마크를 추가할지 여부입니다. (기본값: false) | BOOLEAN | 아니요 | true<br />false |
+
+참고: `model` 옵션에 포함된 텍스트 프롬프트에는 공백이 아닌 문자가 하나 이상 포함되어야 합니다. 이 노드에 표시된 예상 가격은 선택한 비디오 지속 시간을 기준으로 계산됩니다.
+
+## 출력
+
+| 출력 이름 | 설명 | 데이터 타입 |
+|-------------|-------------|-----------|
+| `VIDEO` | 제공된 텍스트 프롬프트에서 생성된 비디오입니다. | VIDEO |
+
+> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03TextToVideoNode/ko.md)
+
+---
+**Source fingerprint (SHA-256):** `9478576dd02ed407a39c95c7227eb8e1482db8b77adc814691fbd807e4cc2893`
diff --git a/zh/built-in-nodes/GeminiVideoOmni.mdx b/zh/built-in-nodes/GeminiVideoOmni.mdx
index ccce9262a..bddab7c1e 100644
--- a/zh/built-in-nodes/GeminiVideoOmni.mdx
+++ b/zh/built-in-nodes/GeminiVideoOmni.mdx
@@ -6,28 +6,30 @@ icon: "circle"
 mode: wide
 ---
 
-使用 Google 的 Gemini Omni Flash 模型，根据文本提示生成带音频的视频。可选择提供参考图像和/或视频来引导或编辑生成结果。请在提示中直接描述所需的时长（3-10 秒）和宽高比（16:9 或 9:16）。
+使用 Google 的 Gemini Omni Flash 模型，根据文本提示生成带音频的视频。可以选择提供参考图片和/或视频来指导或编辑生成结果。请在提示中直接描述所需的时长（3-10秒）和宽高比（16:9 或 9:16）。
 
-## ## 输入
-
-| 参数 | 描述 | 数据类型 | 是否必填 | 范围 |
+## 输入
+| 参数 | 说明 | 数据类型 | 必需 | 范围 |
 |-----------|-------------|-----------|----------|-------|
-| `model` | 用于生成视频的 Gemini 视频模型。 | MODEL | 是 | "Omni Flash" |
-| `seed` | 种子控制节点是否应重新运行；无论种子如何，结果均非确定性（默认值：42）。 | INT | 是 | 0 至 2147483647 |
-| `prompt` | 描述要生成视频的文本提示。 去除开头和结尾的空白字符后，至少需要一个非空白字符。 | STRING | 是 | 去除首尾空格后至少 1 个字符 |
-| `images` | 可选参考图像，用于引导视频生成。最多 14 张图像。 | IMAGE | 否 | 允许多张图像（最多 14 张） |
-| `videos` | 可选参考视频，用于引导或编辑视频生成。最多 3 个视频，每个最长 10 秒。 | VIDEO | 否 | 允许多个视频（最多 3 个，每个最长 10 秒） |
+| `model` | 用于生成视频的 Gemini 视频模型。 | COMBO | 是 | "Omni Flash" |
+| `seed` | 种子控制节点是否应重新运行；无论种子如何，结果都是非确定性的（默认值：42）。 | INT | 是 | 0 至 2147483647 |
+| `prompt` | 描述要生成视频的文本提示。去除首尾空白后，必须至少包含一个非空白字符。 | STRING | 是 | 去除首尾空格后至少 1 个字符 |
+| `images` | 可选的参考图片，用于指导视频生成。总计最多 14 张图片。 | IMAGE | 否 | 允许多张图像（最多 14 张） |
+| `videos` | 可选的参考视频，用于指导或编辑视频生成。最多 3 个视频，每个最长 10 秒。 | VIDEO | 否 | 允许多个视频（最多 3 个，每个最长 10 秒） |
 | `temperature` | 控制生成过程中的随机性（默认值：1.0）。 | FLOAT | 否 | 0.0 至 2.0 |
 | `top_p` | 核采样参数（默认值：0.95）。 | FLOAT | 否 | 0.0 至 1.0 |
 
-## ## 输出
+注意：
+- 如果图片输入包含多帧，则每一帧都计入 14 张图片的上限。
+- 当提供 `images` 或 `videos` 时，编码后媒体的总大小必须保持在大约 90 MB 以下；否则节点会引发错误。
 
-| 输出名称 | 描述 | 数据类型 |
+## 输出
+| 输出名称 | 说明 | 数据类型 |
 |-------------|-------------|-----------|
-| `VIDEO` | 由 Gemini 模型生成的带音频的视频。 | VIDEO |
-| `STRING` | 模型的任何文本响应，例如推理过程或解释。 | STRING |
+| `VIDEO` | 从 Gemini 模型生成的带音频的视频。 | VIDEO |
+| `STRING` | 模型的任何文本响应，例如推理或解释。 | STRING |
 
 > 本文档由 AI 生成。如果您发现任何错误或有改进建议，欢迎贡献！ [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/GeminiVideoOmni/zh.md)
 
 ---
-**Source fingerprint (SHA-256):** `046842b7ec736283bba355aaa038b02fcf2416020f5f7aee7b0150d2a05bcbe6`
+**Source fingerprint (SHA-256):** `1b7ca51d07cfb6a166cfed2a7e7174fd62f3290abcc1bdfdce94369dda242d3f`
diff --git a/zh/built-in-nodes/IdeogramPImage.mdx b/zh/built-in-nodes/IdeogramPImage.mdx
new file mode 100644
index 000000000..4188b1a06
--- /dev/null
+++ b/zh/built-in-nodes/IdeogramPImage.mdx
@@ -0,0 +1,34 @@
+---
+title: "IdeogramPImage - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the IdeogramPImage node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "IdeogramPImage"
+icon: "circle"
+mode: wide
+---
+
+Ideogram P-Image 使用 Ideogram 的高速文本到图像模型，根据文本提示生成图像，该模型以强大的排版和逼真效果著称。它还支持 Ideogram 4.0 结构化 JSON 字幕，可精确控制文本字符串、颜色和布局。此节点返回生成的图像以及实际用于生成图像的最后提示词。
+
+## 输入
+
+| 参数 | 描述 | 数据类型 | 必填 | 范围 |
+|-----------|-------------|-----------|----------|-------|
+| `prompt` | 文本提示。也接受 Ideogram 4.0 结构化 JSON 字幕（精确颜色为 #RRGGBB 十六进制、精确文本字符串、边界框布局）——将 prompt_upsampling 设为 OFF 以按原样使用。不可为空。（默认：""） | STRING | 是 | 任意文本 |
+| `quality` | 速度/价格/质量档位。MEDIUM 是日常默认；HIGH 用于复杂提示、精细细节和困难文本；VERY_LOW/LOW 用于大规模草稿。低于 MEDIUM 时困难文本渲染效果不佳。（默认："MEDIUM"） | STRING | 是 | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
+| `resolution` | 输出尺寸等级（具体像素取决于宽高比，例如 16:9 在 1K 下为 1280x720，在 2K 下为 2560x1440）。若追求清晰排版，建议使用 HIGH + 2K。（默认："1K"） | STRING | 是 | "1K"<br />"2K" |
+| `aspect_ratio` | 图像生成的宽高比。（默认："1:1"） | STRING | 是 | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
+| `prompt_upsampling` | 在生成前将简短提示扩展为详细的结构化字幕（重写后的提示会作为 final_prompt 返回）。在提供自己的 JSON 字幕或精确措辞时设为 OFF。（默认："AUTO"） | STRING | 是 | "AUTO"<br />"ON"<br />"OFF" |
+| `seed` | 用于可复现生成的种子。在 prompt_upsampling 为 OFF 时，相同种子和设置会返回相同图像；在 ON/AUTO 时，提示重写会随每次运行而变化——如需复现结果，可将该图像的 final_prompt 输出与 prompt_upsampling OFF 及相同种子一起再次使用。（默认：42） | INT | 否 | 0 到 2147483647 |
+
+**约束说明：** 提示必须包含至少一个非空白字符，否则节点将失败。在提供自己的结构化 JSON 字幕或精确措辞时，请将 `prompt_upsampling` 设为 OFF。当 `prompt_upsampling` 为 ON 或 AUTO 时，提示会在生成前被重写，因此相同的种子可能无法复现相同的图像；要复现图像，请将 `final_prompt` 输出与 `prompt_upsampling` OFF 及相同种子一起重新使用。
+
+## 输出
+
+| 输出名称 | 描述 | 数据类型 |
+|-------------|-------------|-----------|
+| `image` | 生成的图像，以图像批次形式返回。如果 Ideogram 的内容安全过滤器阻止了生成，则会引发错误。 | IMAGE |
+| `final_prompt` | 实际用于生成图像的提示（如果 prompt_upsampling 运行，则为重写后的结构化字幕，否则为您的提示）。将其与 prompt_upsampling OFF 及相同种子一起反馈回去，即可复现此图像。 | STRING |
+
+> 本文档由 AI 生成。如果您发现任何错误或有改进建议，欢迎贡献！ [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/IdeogramPImage/zh.md)
+
+---
+**Source fingerprint (SHA-256):** `7bd20aae508fee111ded32e87119ed6fc01c5ad5ba7d595e24391830a0f20bb7`
diff --git a/zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx b/zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
new file mode 100644
index 000000000..eaf1d7a53
--- /dev/null
+++ b/zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
@@ -0,0 +1,34 @@
+---
+title: "MinimaxHailuo03FirstLastFrameNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03FirstLastFrameNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03FirstLastFrameNode"
+icon: "circle"
+mode: wide
+---
+
+此节点使用 MiniMax H3 模型，从首帧图像和可选的末帧图像生成视频。视频遵循所提供图像的宽高比，并且当提供末帧时，会从首帧向末帧进行动画生成。
+
+## 输入
+| Parameter | Description | Data Type | Required | Range |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 用于视频生成的模型。此组合包括模型选择（“MiniMax H3”）、描述要生成视频的文本提示、输出分辨率和视频时长。提示必须至少包含一个非空白字符。 | COMBO | 是 | "MiniMax H3" |
+| `first_frame` | 视频的首帧图像。生成视频的宽高比遵循该图像。图像必须至少为 256x256 像素，且宽高比在 0.4 到 2.5 之间。 | IMAGE | 是 | - |
+| `last_frame` | 可选的视频末帧图像。提供后，视频将从首帧开始，以该图像结束。必须满足与 `first_frame` 相同的尺寸和宽高比要求。 | IMAGE | 否 | - |
+| `seed` | 随机种子。使用相同种子的相同请求会产生相似但不保证完全相同的结果。包含“生成后控制”选项，可在每次生成后随机化。默认值：42。 | INT | 是 | 0 到 4294967295 |
+| `watermark` | 是否在视频中添加 AIGC 水印。这是一个高级参数。默认值：False。 | BOOLEAN | 是 | True<br />False |
+
+**约束说明：**
+- `model` 组合中的文本提示不能为空；仅包含空白的提示将被拒绝。
+- 任何提供的帧图像（`first_frame`，以及若使用 `last_frame`）必须至少为 256 像素宽和 256 像素高，宽高比在 0.4 到 2.5 之间（约为 2:5 到 5:2）。
+- `last_frame` 是可选的。省略时，视频仅从首帧生成。
+- 输出视频的宽高比由提供的图像决定，而非由单独的宽高比设置决定。
+
+## 输出
+| Output Name | Description | Data Type |
+|-------------|-------------|-----------|
+| `video` | 使用 MiniMax H3 模型从首帧和可选的末帧生成的视频。 | VIDEO |
+
+> 本文档由 AI 生成。如果您发现任何错误或有改进建议，欢迎贡献！ [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03FirstLastFrameNode/zh.md)
+
+---
+**Source fingerprint (SHA-256):** `f4cb9217eb346019680c64b30c1beacce16f0050616b7b76265edc5840f6b21e`
diff --git a/zh/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx b/zh/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
new file mode 100644
index 000000000..e60033485
--- /dev/null
+++ b/zh/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
@@ -0,0 +1,40 @@
+---
+title: "MinimaxHailuo03ReferenceNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03ReferenceNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03ReferenceNode"
+icon: "circle"
+mode: wide
+---
+
+此节点使用 MiniMax H3 模型生成视频，并通过参考图像、视频和音频来调节结果。在提示词中，参考媒体按其连接顺序引用：“Image 1”、“Image 2”、“Video 1”、“Audio 1”，依此类推。
+
+## 输入
+| 参数 | 描述 | 数据类型 | 必需 | 范围 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 用于视频生成的模型（默认值："MiniMax H3"）。选择"MiniMax H3"将提供下方的 `prompt`、`duration`、`resolution`、`ratio`、`reference_images`、`reference_videos` 和 `reference_audios` 设置。 | STRING | 是 | "MiniMax H3" |
+| `prompt` | 要生成视频的文本描述。参考媒体可按其顺序引用，例如"Image 1"、"Image 2"、"Video 1"或"Audio 1"。 | STRING | 是 | 最小长度：1 个字符 |
+| `duration` | 生成视频的时长（秒）。 | INT | 是 | 多个选项可用 |
+| `resolution` | 生成视频的输出分辨率。 | STRING | 是 | 多个选项可用 |
+| `ratio` | 生成视频的宽高比。 | STRING | 是 | 多个选项可用 |
+| `reference_images` | 主体或风格参考图像，在提示词中按连接顺序称为"Image 1".."Image 9"。最多 9 张图像。 | IMAGE | 否 | 0 到 9 张图像 |
+| `reference_videos` | 运动或场景参考视频，在提示词中按连接顺序称为"Video 1".."Video 3"。最多 3 个视频，每个 2-15 秒，总计 15 秒。 | VIDEO | 否 | 0 到 3 个视频 |
+| `reference_audios` | 音频参考，在提示词中按连接顺序称为"Audio 1".."Audio 3"。最多 3 个片段，每个 2-15 秒，总计 15 秒。不能在没有参考图像或视频的情况下使用。 | AUDIO | 否 | 0 到 3 个片段 |
+| `seed` | 随机种子。相同种子和相同请求会产生相似但不保证完全相同的结果（默认值：42）。 | INT | 是 | 0 到 4294967295 |
+| `watermark` | 是否在视频中添加 AIGC 水印（默认值：false）。 | BOOLEAN | 否 | true<br />false |
+
+### 参数约束
+
+- 至少需要一个参考图像或一个参考视频。仅提供参考音频不被接受。
+- 每个参考图像的宽高比必须约在 0.4 到 2.5（2:5 到 5:2）之间，且最小宽度和高度为 256 像素。
+- 每个参考视频的时长必须在 2 到 15 秒之间，帧率在 23.976 到 60 FPS 之间。所有参考视频的总时长不能超过 15 秒。
+- 每个参考音频片段的时长必须在 2 到 15 秒之间。所有参考音频片段的总时长不能超过 15 秒。
+
+## 输出
+| 输出名 | 描述 | 数据类型 |
+|-------------|-------------|-----------|
+| `video` | 生成的视频。 | VIDEO |
+
+> 本文档由 AI 生成。如果您发现任何错误或有改进建议，欢迎贡献！ [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03ReferenceNode/zh.md)
+
+---
+**Source fingerprint (SHA-256):** `beca020333a544188e6c21829eb8e63415aa5299efc676438e85662a5f08660d`
diff --git a/zh/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx b/zh/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
new file mode 100644
index 000000000..3d3a227b5
--- /dev/null
+++ b/zh/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
@@ -0,0 +1,28 @@
+---
+title: "MinimaxHailuo03TextToVideoNode - ComfyUI Built-in Node Documentation"
+description: "Complete documentation for the MinimaxHailuo03TextToVideoNode node in ComfyUI. Learn its inputs, outputs, parameters and usage."
+sidebarTitle: "MinimaxHailuo03TextToVideoNode"
+icon: "circle"
+mode: wide
+---
+
+此节点使用 MiniMax H3 模型根据文本提示生成视频。它将文本以及分辨率、时长和宽高比等视频设置发送到 MiniMax API，并返回生成的视频作为输出。
+
+## 输入
+| 参数 | 描述 | 数据类型 | 是否必需 | 范围 |
+|-----------|-------------|-----------|----------|-------|
+| `model` | 用于视频生成的模型。（默认值："MiniMax H3"）。此选择还包括生成视频的文本提示、分辨率、时长和宽高比设置。 | COMBO | 是 | `"MiniMax H3"` |
+| `seed` | 随机种子。使用相同种子的相同请求会得到相似但不保证完全相同的结果。（默认值：42） | INT | 是 | 0 到 4294967295 |
+| `watermark` | 是否向视频添加 AIGC 水印。（默认值：false） | BOOLEAN | 否 | true<br />false |
+
+注意：`model` 选项中包含的文本提示必须至少包含一个非空白字符。此节点显示的预估价格根据所选视频时长计算。
+
+## 输出
+| 输出名称 | 描述 | 数据类型 |
+|-------------|-------------|-----------|
+| `VIDEO` | 根据所提供文本提示生成的视频。 | VIDEO |
+
+> 本文档由 AI 生成。如果您发现任何错误或有改进建议，欢迎贡献！ [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/MinimaxHailuo03TextToVideoNode/zh.md)
+
+---
+**Source fingerprint (SHA-256):** `9478576dd02ed407a39c95c7227eb8e1482db8b77adc814691fbd807e4cc2893`
diff --git a/zh/built-in-nodes/RunwayFirstLastFrameNode.mdx b/zh/built-in-nodes/RunwayFirstLastFrameNode.mdx
index d1088069e..fce13e32a 100644
--- a/zh/built-in-nodes/RunwayFirstLastFrameNode.mdx
+++ b/zh/built-in-nodes/RunwayFirstLastFrameNode.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **已退役（EOL）**：此节点已于 **2026 年 7 月 30 日** 达到生命周期终点。首尾帧视频请使用 [Kling 3.0 首尾帧转视频](/zh/tutorials/partner-nodes/kling/kling-3-0) 或其他受支持的 FLF2V 模板。
-</Warning>
-
-Runway 首尾帧生成视频节点通过上传首帧、尾帧及文本提示词来生成视频。
+Runway 首尾帧生成视频节点通过上传首帧、尾帧及文本提示词来生成视频。该节点利用 Runway 的 Gen-3 模型，在提供的起始帧与结束帧之间创建平滑过渡。当结束帧与起始帧差异显著时，此功能尤其适用于复杂转场场景。
 
 ## 输入
 
diff --git a/zh/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx b/zh/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
index 780e2a056..816c11381 100644
--- a/zh/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
+++ b/zh/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
@@ -6,13 +6,7 @@ icon: "circle"
 mode: wide
 ---
 
----
-
-<Warning>
-  **已退役（EOL）**：此节点已于 **2026 年 7 月 30 日** 达到生命周期终点（Runway 下线 Gen-3 Alpha Turbo）。请改用 [Runway Gen4 Turbo](/zh/built-in-nodes/RunwayImageToVideoNodeGen4) 或 [Gen4 Turbo 工作流](/zh/tutorials/partner-nodes/runway/video-generation)。
-</Warning>
-
-Runway 图像转视频（Gen3a Turbo）节点使用 Runway 的 Gen3a Turbo 模型，从单个起始帧生成视频。
+Runway 图像转视频（Gen3a Turbo）节点使用 Runway 的 Gen3a Turbo 模型，从单个起始帧生成视频。它接收文本提示和初始图像帧，然后根据指定的时长和宽高比创建视频序列。此节点连接到 Runway 的 API 以远程处理生成过程。
 
 ## 输入