OpenAI-Compatible Models
OpenAI-Compatible Models
In addition to Google Gemini, AMC WebUI connects to standard OpenAI /chat/completions API endpoints across the AI ecosystem.
Configuration Isolation
AMC WebUI implements strict data isolation:
- Independent Base URL: Configured in Settings -> Providers & API -> OpenAI Compatible, completely separated from Gemini routes.
- Independent Key Storage: Saved in a separate IndexedDB table, preventing accidental exposure to Google endpoints.
- Custom Model Lists: Manage custom model names (e.g.,
deepseek-chat,deepseek-reasoner,claude-3-5-sonnet) independently.
Provider Setup Examples
1. DeepSeek
- Base URL:
https://api.deepseek.com/v1 - Models:
deepseek-chat,deepseek-reasoner - Reasoning Support: AMC WebUI detects
deepseek-reasonerand folds thinking tokens into the UI thought container.
2. OpenRouter
- Base URL:
https://openrouter.ai/api/v1 - Models:
anthropic/claude-3.5-sonnet,meta-llama/llama-3.3-70b-instruct
3. Google AI Studio (OpenAI Endpoint)
- Base URL:
https://generativelanguage.googleapis.com/v1beta/openai - API Key: Use your
GEMINI_API_KEY
Server-Side Proxy Routing (THIRD_PARTY_ROUTES)
When deploying via Docker, configure THIRD_PARTY_ROUTES to route requests through your server proxy:
{ "deepseek": { "baseUrl": "https://api.deepseek.com/v1", "apiKey": "sk-your-server-deepseek-key" }}- Requests go to
/api/openai/*with headerx-third-party-provider: deepseek. - Browser API keys take priority, falling back to server keys when absent.