CineCapture builds open-source tools for recording and directing cinematic video inside Unity applications at runtime.

The projects cover the complete capture pipeline: camera control, Unity frame capture, GPU video encoding, audio/video muxing, and sample integrations.

Use one camera, record several cameras independently, or combine multiple sources into one automatically edited video.

Follow the UnityRuntimeCameraRecorder Getting Started guide to integrate cameras, screen capture, textures, multi-source sequences, transitions, statistics and image-sequence capture step by step.

These videos were recorded with the UnitySample application. The first three runs saved each selected camera as a separate video. The fourth run combined three selected sources into one video.

One camera was recorded with the High preset.

Camera2

Screen Camera and Camera1 were recorded simultaneously with the High preset.

Screen Camera	Camera1

Screen Camera and Camera1 were recorded simultaneously with the Low preset.

Screen Camera	Camera1

Camera1, Camera2, and Screen Camera were recorded into one automatically edited video.

Camera1 + Camera2 + Screen Camera → 1 output

These measurements used an NVIDIA GeForce RTX 5060, H.264, 4K at 60 FPS, VSync, and MSAA 4x. The two-camera runs recorded Camera 1 and Screen as separate videos; the single-camera run recorded Camera 2. The fourth run combined Camera 1, Camera 2, and Screen Camera into one output. Video FPS was measured from second 1 through second 6.

All configurations maintained essentially 60 FPS. Low reduced the combined two-camera output size by about 70%.

flowchart TB

    subgraph Sample["UnitySample — 2 cameras recording example"]
        direction LR
        Camera1["Camera 1<br/>Scene view"]
        Camera2["Camera 2<br/>Alternate view"]
        Audio["Unity audio mix"]
    end

    Camera1 --> Recorder1["UnityRuntimeCameraRecorder #1"]
    Camera2 --> Recorder2["UnityRuntimeCameraRecorder #2"]

    Audio --> Recorder1
    Audio --> Recorder2

    Recorder1 --> Encoder1["Direct3DVideoEncoder"]
    Recorder2 --> Encoder2["Direct3DVideoEncoder"]

    Encoder1 --> Writer1["FFmpegMediaWriter"]
    Encoder2 --> Writer2["FFmpegMediaWriter"]

    Recorder1 -->|"PCM audio"| Writer1
    Recorder2 -->|"PCM audio"| Writer2

    Writer1 --> Output1["Camera 1.mp4"]
    Writer2 --> Output2["Camera 2.mp4"]

    classDef camera fill:#593d88,color:#fff,stroke:#b99ae8,stroke-width:2px;
    classDef library fill:#176b87,color:#fff,stroke:#64ccc5,stroke-width:2px;
    classDef audio fill:#2d4356,color:#fff,stroke:#a7c4bc,stroke-width:2px;
    classDef output fill:#1f8a70,color:#fff,stroke:#9de8d7,stroke-width:2px;
    class Camera1,Camera2 camera;
    class Recorder1,Recorder2,Encoder1,Encoder2,Writer1,Writer2 library;
    class Audio audio;
    class Output1,Output2 output;

Want to help move things off the wishlist? Contributions are more than welcome!

You can help by testing the recorder on your OS and GPU, then sharing your setup and performance results. Please include your OS, GPU model, Unity version, graphics API, and anything that worked—or didn't.

Feeling like writing some code? You can also help build support for the platforms, GPUs, and graphics APIs that aren't supported yet. Pick something from the wishlist, open an issue to discuss your approach, and let's make it happen.