Chat apps: Open WebUI, Cherry Studio, LobeChat

All three speak the OpenAI format. Use one 666 AI key and pick any model.

Replace YOUR-ENDPOINT with the endpoint shown on your API Key page after you sign up, and YOUR-KEY with your key. One key works for every model on 666 AI — Claude, GPT, Gemini, DeepSeek, Qwen and 200+ more.

Open WebUI

  1. Settings → Admin → Connections → Manage OpenAI API Connections → Add Connection.
  2. URL: https://YOUR-ENDPOINT/v1, API Key: YOUR-KEY.
  3. Optional: add only the models you want under Model IDs (e.g. anthropic/claude-sonnet-5.5) — otherwise it loads the full list.
  4. Verify Connection, then Save.

Or with environment variables (first start only — after that, change it in the Admin UI):

OPENAI_API_BASE_URL=https://YOUR-ENDPOINT/v1
OPENAI_API_KEY=YOUR-KEY

Cherry Studio

  1. Settings → Model Providers → + Add, type OpenAI.
  2. API Host: https://YOUR-ENDPOINT (Cherry Studio adds /v1 itself; with /v1 also works).
  3. API Key: YOUR-KEY.
  4. Fetch the model list or add IDs by hand (e.g. google/gemini-3.5-flash), then Check and switch the provider on.

LobeChat / LobeHub

  1. Settings → Provider → Add Custom Provider; Request Format openai.
  2. Proxy URL: https://YOUR-ENDPOINT/v1; API Key: YOUR-KEY.
  3. Fetch models, or Add Model with the exact ID.

Self-hosted with environment variables:

OPENAI_API_KEY=YOUR-KEY
OPENAI_PROXY_URL=https://YOUR-ENDPOINT/v1
OPENAI_MODEL_LIST=-all,+anthropic/claude-sonnet-5.5=Claude Sonnet 5.5,+google/gemini-3.5-flash=Gemini 3.5 Flash

FAQ

Which model should I start with?

For everyday chat google/gemini-3.5-flash is fast and cheap; for hard reasoning and coding, anthropic/claude-sonnet-5.5.

The model list is huge

666 AI exposes about 600 IDs. Add only the ones you use (Model IDs in Open WebUI, manual add in Cherry Studio / LobeChat).

Create a free accountAll modelsDeveloper guides