Lobe Chat

Lobe Chat is an open-source, self-hostable AI chat application. There are two ways to connect it: one-click configuration via URL parameters, or manual setup in the UI.

Option 1: One-click configuration via URL parameters

Lobe Chat supports encoding settings into the access URL. In the link below, replace {url} with your Lobe Chat deployment address and {key} with your FIM Gate key, then open it in a browser:

https://{url}/?settings={"keyVaults":{"openai":{"apiKey":"{key}","baseURL":"https://api.gate.fim.ai/v1"}},"languageModel":{"openai":{"autoFetchModelLists":true,"enabled":true},"ollama":{"enabled":false}},"check_updates":false}

You can also add an array of model IDs under languageModel.openai.enabledModels to pre-enable specific models (e.g. ["gpt-5.2","claude-sonnet-4-5-20250929"]). If omitted, you can enable models in the UI after opening the app.

Option 2: Manual configuration

1. Open Lobe Chat, go to App Settings → Language Model, and find the OpenAI provider.

2. For API Key, enter a key you created in the Console.

3. For API Proxy Address (BASE URL), enter:

   https://api.gate.fim.ai/v1

4. Click Fetch Model List to retrieve the currently available models.

5. In the Model List, check the models you want to use, then return to the chat view to select them.