Improve GLM 4.x provider support

CLAUDE.md

@@ -120,6 +120,7 @@ python test_security_integration.py   # Integration tests (9 tests)
 python -m pytest test_client.py       # Client tests (20 tests)
 python -m pytest test_dependency_resolver.py  # Dependency resolver tests (12 tests)
 python -m pytest test_rate_limit_utils.py     # Rate limit tests (22 tests)
+python -m pytest test_api_providers.py        # API provider env tests
 ```
 
 ### React UI
@@ -424,6 +425,11 @@ Alternative providers are configured via the **Settings UI** (gear icon > API Pr
 
 **Available providers:** Claude (default), GLM (Zhipu AI), Ollama (local models), Kimi (Moonshot), Custom
 
+**GLM notes:**
+- Auth via `ANTHROPIC_AUTH_TOKEN`; base URL defaults to `https://api.z.ai/api/anthropic` (editable, e.g. `https://open.bigmodel.cn/api/anthropic` for mainland China)
+- Models: `glm-5`, `glm-4.7` (default), `glm-4.6`, `glm-4.5`, `glm-4.5-air`
+- `API_TIMEOUT_MS` defaults to 3000000 via the provider's `default_timeout_ms` in `registry.py`
+
 **Ollama notes:**
 - Requires Ollama v0.14.0+ with Anthropic API compatibility
 - Install: https://ollama.com → `ollama serve` → `ollama pull qwen3-coder`

README.md

@@ -338,6 +338,8 @@ Alternative providers are configured via the **Settings UI** (gear icon > API Pr
 
 Available providers: **Claude** (default), **GLM** (Zhipu AI), **Ollama** (local models), **Kimi** (Moonshot), **Custom**
 
+**GLM (Zhipu AI / Z.ai):** pick "GLM" in Settings, paste your Z.ai API key (e.g. from the GLM Coding Plan), and choose a model (GLM 5, GLM 4.7, 4.6, 4.5, 4.5 Air, or type any other model ID). The base URL defaults to `https://api.z.ai/api/anthropic`; mainland China accounts should change it to `https://open.bigmodel.cn/api/anthropic`. AutoForge applies Z.ai's recommended `API_TIMEOUT_MS=3000000` automatically so long planning turns don't time out.
+
 For Ollama, install [Ollama v0.14.0+](https://ollama.com), run `ollama serve`, and pull a coding model (e.g., `ollama pull qwen3-coder`). Then select "Ollama" in the Settings UI.
 
 ### Using Vertex AI

registry.py

@@ -801,9 +801,14 @@ def get_all_settings() -> dict[str, str]:
         "models": [
             {"id": "glm-5", "name": "GLM 5"},
             {"id": "glm-4.7", "name": "GLM 4.7"},
+            {"id": "glm-4.6", "name": "GLM 4.6"},
+            {"id": "glm-4.5", "name": "GLM 4.5"},
             {"id": "glm-4.5-air", "name": "GLM 4.5 Air"},
         ],
         "default_model": "glm-4.7",
+        # Z.ai recommends a long timeout for its Anthropic-compatible endpoint;
+        # long planning turns otherwise hit the CLI's default request timeout.
+        "default_timeout_ms": "3000000",
     },
     "azure": {
         "name": "Azure Anthropic (Claude)",
@@ -907,7 +912,7 @@ def get_effective_sdk_env() -> dict[str, str]:
         sdk_env["ANTHROPIC_DEFAULT_HAIKU_MODEL"] = model
 
     # Timeout
-    timeout = all_settings.get("api_timeout_ms")
+    timeout = all_settings.get("api_timeout_ms") or provider.get("default_timeout_ms")
     if timeout:
         sdk_env["API_TIMEOUT_MS"] = timeout
 

test_api_providers.py

@@ -0,0 +1,49 @@
+"""Tests for alternative API provider env construction (registry.get_effective_sdk_env)."""
+
+import pytest
+
+import registry
+
+
+@pytest.fixture
+def settings(monkeypatch):
+    store: dict[str, str] = {}
+    monkeypatch.setattr(registry, "get_all_settings", lambda: dict(store))
+    return store
+
+
+def test_glm_defaults(settings):
+    settings.update({"api_provider": "glm", "api_auth_token": "zai-key"})
+    env = registry.get_effective_sdk_env()
+
+    assert env["ANTHROPIC_BASE_URL"] == "https://api.z.ai/api/anthropic"
+    assert env["ANTHROPIC_AUTH_TOKEN"] == "zai-key"
+    assert env["ANTHROPIC_API_KEY"] == ""
+    assert env["API_TIMEOUT_MS"] == "3000000"
+    for tier in ("OPUS", "SONNET", "HAIKU"):
+        assert env[f"ANTHROPIC_DEFAULT_{tier}_MODEL"] == "glm-4.7"
+
+
+def test_glm_overrides(settings):
+    settings.update({
+        "api_provider": "glm",
+        "api_auth_token": "key",
+        "api_base_url": "https://open.bigmodel.cn/api/anthropic",
+        "api_model": "glm-4.6",
+        "api_timeout_ms": "600000",
+    })
+    env = registry.get_effective_sdk_env()
+
+    assert env["ANTHROPIC_BASE_URL"] == "https://open.bigmodel.cn/api/anthropic"
+    assert env["ANTHROPIC_DEFAULT_SONNET_MODEL"] == "glm-4.6"
+    assert env["API_TIMEOUT_MS"] == "600000"
+
+
+def test_glm_lists_4x_models():
+    ids = {m["id"] for m in registry.API_PROVIDERS["glm"]["models"]}
+    assert {"glm-4.7", "glm-4.6", "glm-4.5", "glm-4.5-air"} <= ids
+
+
+def test_providers_without_default_timeout_leave_it_unset(settings):
+    settings.update({"api_provider": "kimi", "api_auth_token": "k"})
+    assert "API_TIMEOUT_MS" not in registry.get_effective_sdk_env()

ui/src/components/SettingsModal.tsx

@@ -23,7 +23,7 @@ interface SettingsModalProps {
 const PROVIDER_INFO_TEXT: Record<string, string> = {
   claude: 'Default provider. Uses Claude CLI credentials. API key auth is recommended.',
   kimi: 'Get an API key at kimi.com',
-  glm: 'Get an API key at open.bigmodel.cn',
+  glm: 'Get an API key at z.ai (GLM Coding Plan). Mainland China accounts: change the Base URL to https://open.bigmodel.cn/api/anthropic',
   ollama: 'Run models locally. Install from ollama.com',
   custom: 'Connect to any OpenAI-compatible API endpoint.',
 }
@@ -116,7 +116,7 @@ export function SettingsModal({ isOpen, onClose }: SettingsModalProps) {
   const currentProviderInfo: ProviderInfo | undefined = providers.find(p => p.id === currentProvider)
   const isAlternativeProvider = currentProvider !== 'claude'
   const showAuthField = isAlternativeProvider && currentProviderInfo?.requires_auth
-  const showBaseUrlField = currentProvider === 'custom' || currentProvider === 'azure'
+  const showBaseUrlField = currentProvider === 'custom' || currentProvider === 'azure' || currentProvider === 'glm'
   const showCustomModelInput = currentProvider === 'custom' || currentProvider === 'ollama'
 
   return (
@@ -332,7 +332,7 @@ export function SettingsModal({ isOpen, onClose }: SettingsModalProps) {
                         type="text"
                         value={customBaseUrlInput}
                         onChange={(e) => setCustomBaseUrlInput(e.target.value)}
-                        placeholder={currentProvider === 'azure' ? 'https://your-resource.services.ai.azure.com/anthropic' : 'https://api.example.com/v1'}
+                        placeholder={currentProvider === 'azure' ? 'https://your-resource.services.ai.azure.com/anthropic' : currentProvider === 'glm' ? 'https://api.z.ai/api/anthropic' : 'https://api.example.com/v1'}
                         className="flex-1 py-1.5 px-3 text-sm border rounded-md bg-background"
                       />
                       <Button

ui/src/components/views/SettingsView.tsx

@@ -21,7 +21,7 @@ import { Card, CardContent, CardHeader, CardTitle } from '@/components/ui/card'
 const PROVIDER_INFO_TEXT: Record<string, string> = {
   claude: 'Default provider. Uses Claude CLI credentials. API key auth is recommended.',
   kimi: 'Get an API key at kimi.com',
-  glm: 'Get an API key at open.bigmodel.cn',
+  glm: 'Get an API key at z.ai (GLM Coding Plan). Mainland China accounts: change the Base URL to https://open.bigmodel.cn/api/anthropic',
   ollama: 'Run models locally. Install from ollama.com',
   custom: 'Connect to any OpenAI-compatible API endpoint.',
 }
@@ -113,7 +113,7 @@ export function SettingsView() {
   const currentProviderInfo: ProviderInfo | undefined = providers.find(p => p.id === currentProvider)
   const isAlternativeProvider = currentProvider !== 'claude'
   const showAuthField = isAlternativeProvider && currentProviderInfo?.requires_auth
-  const showBaseUrlField = currentProvider === 'custom' || currentProvider === 'azure'
+  const showBaseUrlField = currentProvider === 'custom' || currentProvider === 'azure' || currentProvider === 'glm'
   const showCustomModelInput = currentProvider === 'custom' || currentProvider === 'ollama'
 
   return (
@@ -341,7 +341,7 @@ export function SettingsView() {
                             type="text"
                             value={customBaseUrlInput}
                             onChange={(e) => setCustomBaseUrlInput(e.target.value)}
-                            placeholder={currentProvider === 'azure' ? 'https://your-resource.services.ai.azure.com/anthropic' : 'https://api.example.com/v1'}
+                            placeholder={currentProvider === 'azure' ? 'https://your-resource.services.ai.azure.com/anthropic' : currentProvider === 'glm' ? 'https://api.z.ai/api/anthropic' : 'https://api.example.com/v1'}
                             className="flex-1 py-1.5 px-3 text-sm border rounded-md bg-background"
                           />
                           <Button