feat: add GLM (Z.ai) LLM provider

docs/en_US/ai_tools.rst

@@ -26,7 +26,7 @@ Before using AI Reports, you must:
 
 **Note:**
 
- * AI Reports using cloud providers (Anthropic, OpenAI) require an active internet connection.
+ * AI Reports using cloud providers (Anthropic, OpenAI, GLM) require an active internet connection.
    Local providers (Ollama, Docker Model Runner) do not require internet access.
 
  * API usage may incur costs depending on your LLM provider's pricing model.
@@ -62,6 +62,13 @@ Select your preferred LLM provider from the dropdown:
   * **API Key File**: Path to a file containing your OpenAI API key (obtain from https://platform.openai.com/). This path refers to the filesystem where the pgAdmin server is running (e.g., inside the container if using Docker). The ``~`` prefix is expanded to the home directory of the user running the pgAdmin server process. Optional when using a custom URL with a provider that does not require authentication.
   * **Model**: Select from available GPT models (e.g., gpt-4).
 
+**GLM (Z.ai)**
+  Use GLM models from Z.ai through the provider's OpenAI-compatible API.
+
+  * **API URL**: Custom API endpoint URL (leave empty for default: https://api.z.ai/api/paas/v4). Use https://api.z.ai/api/coding/paas/v4 with a Z.ai Coding Plan subscription.
+  * **API Key File**: Path to a file containing your Z.ai API key. This path refers to the filesystem where the pgAdmin server is running (e.g., inside the container if using Docker). The ``~`` prefix is expanded to the home directory of the user running the pgAdmin server process.
+  * **Model**: Select from available GLM models (e.g., glm-4.7). There is no default; a model must be selected before the AI features can be used.
+
 **Ollama**
   Use locally-hosted open-source models via Ollama. Requires a running Ollama instance.
 
@@ -237,7 +244,7 @@ Troubleshooting
   You must establish a connection to the server or database before generating reports.
 
 **API Connection Errors**
-  * Verify your API key is correct (for Anthropic and OpenAI).
+  * Verify your API key is correct (for Anthropic, OpenAI, and GLM).
   * Check your internet connection (for cloud providers).
   * For Ollama, ensure the Ollama server is running and accessible.
   * For Docker Model Runner, ensure Docker Desktop 4.40+ is running with the model runner enabled.

docs/en_US/preferences.rst

@@ -43,7 +43,7 @@ in ``config.py``) for these preferences to be available.
 Use the fields on the *AI* panel to configure your LLM provider:
 
 * Use the *Default Provider* drop-down to select your LLM provider. Options include:
-  *Anthropic*, *OpenAI*, *Ollama*, or *Docker Model Runner*.
+  *Anthropic*, *OpenAI*, *GLM (Z.ai)*, *Ollama*, or *Docker Model Runner*.
 
 **Anthropic Settings:**
 
@@ -81,6 +81,24 @@ Use the fields on the *AI* panel to configure your LLM provider:
   refresh button to fetch the latest available models from your configured
   endpoint.
 
+**GLM (Z.ai) Settings:**
+
+* Use the *API URL* field to set a custom API endpoint URL. Leave empty to use
+  the default GLM API (``https://api.z.ai/api/paas/v4``). Use
+  ``https://api.z.ai/api/coding/paas/v4`` with a Z.ai Coding Plan
+  subscription.
+
+* Use the *API Key File* field to specify the path to a file containing your
+  GLM API key. This path refers to the filesystem where the pgAdmin
+  server is running (e.g., inside the container if using Docker). The ``~``
+  prefix is expanded to the home directory of the user running the pgAdmin
+  server process.
+
+* Use the *Model* field to select from the available GLM models. Click the
+  refresh button to fetch the latest available models from your configured
+  endpoint. There is no default model, so one must be selected before the
+  AI features can be used.
+
 **Ollama Settings:**
 
 * Use the *API URL* field to specify the Ollama server URL

web/config.py

@@ -1025,7 +1025,8 @@
 # Default LLM Provider
 # Specifies which LLM provider to use by default when LLM_ENABLED is True.
 # Users can override this in their preferences.
-# Valid values: 'anthropic', 'openai', 'ollama', 'docker', or '' (disabled)
+# Valid values: 'anthropic', 'openai', 'glm', 'ollama', 'docker',
+# or '' (disabled)
 DEFAULT_LLM_PROVIDER = ''
 
 # Anthropic Configuration
@@ -1064,6 +1065,24 @@
 # Examples: gpt-4o, gpt-4o-mini, gpt-4-turbo
 OPENAI_API_MODEL = ''
 
+# GLM (Z.ai) Configuration
+# URL for the GLM API endpoint. Leave empty to use the default
+# (https://api.z.ai/api/paas/v4). Set this to
+# https://api.z.ai/api/coding/paas/v4 for a Z.ai Coding Plan
+GLM_API_URL = ''
+
+# Path to a file containing the GLM API key. The file should contain
+# only the API key with no additional whitespace or formatting.
+# Default: ~/.zai-api-key
+GLM_API_KEY_FILE = '~/.zai-api-key'
+
+# The GLM model to use for AI features. There is no default: the models
+# available differ between the Z.ai open platform, Coding Plan
+# subscriptions and Zhipu's China platform, so one must be selected
+# here or in the user's preferences.
+# Examples: glm-4.6, glm-4.7, glm-5
+GLM_API_MODEL = ''
+
 # Ollama Configuration
 # URL for the Ollama API endpoint. Leave empty to disable Ollama.
 # Typical value: http://localhost:11434
@@ -1105,6 +1124,7 @@
 ALLOWED_LLM_API_URLS = [
     'https://api.anthropic.com:443',
     'https://api.openai.com:443',
+    'https://api.z.ai:443',
     # Loopback addresses on any port: covers Ollama (11434), Docker
     # Model Runner (12434), LiteLLM (4000), vLLM (8000), LM Studio
     # (1234), text-generation-webui (5000), and any self-hosted

web/pgadmin/llm/README.md

@@ -13,15 +13,17 @@ This module provides AI/LLM functionality for pgAdmin, including database securi
 
 - **Anthropic Claude** (recommended)
 - **OpenAI GPT**
+- **GLM (Z.ai)**
 - **Ollama** (local models)
 
 ## Configuration
 
 Configure LLM providers in `config.py`:
 
-- `DEFAULT_LLM_PROVIDER`: Set to 'anthropic', 'openai', or 'ollama'
+- `DEFAULT_LLM_PROVIDER`: Set to 'anthropic', 'openai', 'glm', or 'ollama'
 - `ANTHROPIC_API_KEY_FILE`: Path to file containing Anthropic API key
 - `OPENAI_API_KEY_FILE`: Path to file containing OpenAI API key
+- `GLM_API_KEY_FILE`: Path to file containing GLM API key
 - `OLLAMA_API_URL`: URL for Ollama server (e.g., 'http://localhost:11434')
 
 If API keys are not found, the LLM features will be gracefully disabled.

web/pgadmin/llm/__init__.py

@@ -40,7 +40,7 @@
 MODULE_NAME = 'llm'
 
 # Valid LLM providers
-LLM_PROVIDERS = ['anthropic', 'openai', 'ollama', 'docker']
+LLM_PROVIDERS = ['anthropic', 'openai', 'glm', 'ollama', 'docker']
 
 
 class LLMModule(PgAdminModule):
@@ -71,6 +71,7 @@ def register_preferences(self):
             {'label': gettext('None (Disabled)'), 'value': ''},
             {'label': gettext('Anthropic'), 'value': 'anthropic'},
             {'label': gettext('OpenAI'), 'value': 'openai'},
+            {'label': gettext('GLM (Z.ai)'), 'value': 'glm'},
             {'label': gettext('Ollama'), 'value': 'ollama'},
             {'label': gettext('Docker Model Runner'), 'value': 'docker'},
         ]
@@ -238,6 +239,66 @@ def register_preferences(self):
             }
         )
 
+        # GLM Settings
+        # Get defaults from config
+        glm_url_default = getattr(config, 'GLM_API_URL', '')
+        glm_key_file_default = getattr(config, 'GLM_API_KEY_FILE', '')
+        glm_model_default = getattr(config, 'GLM_API_MODEL', '')
+
+        self.glm_api_url = self.preference.register(
+            'glm', 'glm_api_url',
+            gettext("API URL"), 'text',
+            glm_url_default,
+            category_label=gettext('GLM (Z.ai)'),
+            help_str=gettext(
+                'URL for the GLM API endpoint. Leave empty to use '
+                'the default (https://api.z.ai/api/paas/v4). Use '
+                'https://api.z.ai/api/coding/paas/v4 for a Coding Plan '
+                'subscription, or set a custom URL to use another '
+                'GLM-compatible API provider.'
+            )
+        )
+
+        self.glm_api_key_file = self.preference.register(
+            'glm', 'glm_api_key_file',
+            gettext("API Key File"), 'text',
+            glm_key_file_default,
+            category_label=gettext('GLM (Z.ai)'),
+            help_str=gettext(
+                'Path to a file containing your GLM API key. '
+                'This path must be on the server hosting pgAdmin, '
+                'e.g. inside the container when using Docker. '
+                'The file should contain only the API key.'
+            )
+        )
+
+        glm_model_options = []
+
+        self.glm_api_model = self.preference.register(
+            'glm', 'glm_api_model',
+            gettext("Model"), 'options',
+            glm_model_default,
+            category_label=gettext('GLM (Z.ai)'),
+            options=glm_model_options,
+            help_str=gettext(
+                'The GLM model to use. Models are loaded dynamically '
+                'from your API key. You can also type a custom model name. '
+                'A model must be selected; there is no default.'
+            ),
+            control_props={
+                'allowClear': True,
+                'creatable': True,
+                'tags': True,
+                'placeholder': gettext('Select or type a model name...'),
+                'optionsUrl': 'llm.models_glm',
+                'optionsRefreshUrl': 'llm.refresh_models_glm',
+                'refreshDepNames': {
+                    'api_url': 'glm_api_url',
+                    'api_key_file': 'glm_api_key_file'
+                }
+            }
+        )
+
         # Ollama Settings
         # Get defaults from config
         ollama_url_default = getattr(config, 'OLLAMA_API_URL', '')
@@ -334,10 +395,12 @@ def get_exposed_url_endpoints(self):
         return [
             'llm.models_anthropic',
             'llm.models_openai',
+            'llm.models_glm',
             'llm.models_ollama',
             'llm.models_docker',
             'llm.refresh_models_anthropic',
             'llm.refresh_models_openai',
+            'llm.refresh_models_glm',
             'llm.refresh_models_ollama',
             'llm.refresh_models_docker',
             'llm.status',
@@ -378,7 +441,7 @@ def get_llm_status():
     """
     from pgadmin.llm.utils import (
         is_llm_enabled, is_llm_enabled_system, get_default_provider,
-        get_anthropic_model, get_openai_model, get_ollama_model,
+        get_anthropic_model, get_openai_model, get_glm_model, get_ollama_model,
         get_docker_model
     )
 
@@ -388,6 +451,8 @@ def get_llm_status():
         model = get_anthropic_model()
     elif provider == 'openai':
         model = get_openai_model()
+    elif provider == 'glm':
+        model = get_glm_model()
     elif provider == 'ollama':
         model = get_ollama_model()
     elif provider == 'docker':
@@ -617,6 +682,108 @@ def refresh_openai_models():
         )
 
 
+@blueprint.route("/models/glm", methods=["GET"], endpoint='models_glm')
+@pga_login_required
+def get_glm_models():
+    """
+    Fetch available GLM models.
+    Returns models from the OpenAI-compatible models endpoint.
+    """
+    from pgadmin.llm.utils import get_glm_api_key, get_glm_api_url
+
+    api_key = get_glm_api_key()
+    api_url = get_glm_api_url()
+    if not api_key and not api_url:
+        return make_json_response(
+            data={'models': [], 'error': 'No API key configured'},
+            status=200
+        )
+
+    try:
+        models = _fetch_glm_models(api_key, api_url)
+        return make_json_response(data={'models': models}, status=200)
+    except LLMApiError as e:
+        return make_json_response(
+            data={'models': [], 'error': str(e)},
+            status=200
+        )
+    except Exception:
+        return make_json_response(
+            data={'models': [],
+                  'error': 'Failed to fetch GLM models. '
+                           'Check the API key and URL configuration.'},
+            status=200
+        )
+
+
+@blueprint.route(
+    "/models/glm/refresh",
+    methods=["POST"],
+    endpoint='refresh_models_glm'
+)
+@pga_login_required
+def refresh_glm_models():
+    """
+    Fetch available GLM models using a provided API key file path
+    and/or custom API URL.
+    Used by the preferences refresh button to load models before saving.
+    """
+    from pgadmin.llm.utils import (
+        read_api_key_file, validate_api_key_path, validate_api_url
+    )
+
+    data = request.get_json(force=True, silent=True) or {}
+    api_key_file = data.get('api_key_file', '')
+    api_url = data.get('api_url', '')
+
+    if api_url and not validate_api_url(api_url):
+        return make_json_response(
+            data={'models': [],
+                  'error': 'API URL is not in the allowed list. '
+                           'Contact your administrator to update '
+                           'ALLOWED_LLM_API_URLS in the server '
+                           'configuration.'},
+            status=200
+        )
+
+    api_key = None
+    if api_key_file:
+        safe_path = validate_api_key_path(api_key_file)
+        if safe_path is None:
+            return make_json_response(
+                data={'models': [],
+                      'error': 'API key file path is not permitted. '
+                               'The file must be within your private '
+                               'user storage directory; shared storage '
+                               'and other locations are not allowed for '
+                               'security reasons.'},
+                status=200
+            )
+        api_key = read_api_key_file(safe_path)
+
+    if not api_key and not api_url:
+        return make_json_response(
+            data={'models': [], 'error': 'No API key or custom URL provided'},
+            status=200
+        )
+
+    try:
+        models = _fetch_glm_models(api_key, api_url)
+        return make_json_response(data={'models': models}, status=200)
+    except LLMApiError as e:
+        return make_json_response(
+            data={'models': [], 'error': str(e)},
+            status=200
+        )
+    except Exception:
+        return make_json_response(
+            data={'models': [],
+                  'error': 'Failed to fetch GLM models. '
+                           'Check the API key file and URL.'},
+            status=200
+        )
+
+
 @blueprint.route("/models/ollama", methods=["GET"], endpoint='models_ollama')
 @pga_login_required
 def get_ollama_models():
@@ -924,6 +1091,19 @@ def _fetch_openai_models(api_key, api_url=''):
     return models
 
 
+def _fetch_glm_models(api_key, api_url=''):
+    """
+    Fetch models from GLM (Z.ai) API.
+    Returns a list of model options with label and value.
+    """
+    from pgadmin.llm.providers.glm import DEFAULT_API_BASE_URL
+
+    return _fetch_openai_models(
+        api_key=api_key,
+        api_url=api_url or DEFAULT_API_BASE_URL
+    )
+
+
 def _fetch_ollama_models(api_url):
     """
     Fetch models from Ollama API.