Skip to content
RunAPI Developer Docs
EN 中文
Developer Resources
Developer Resources

Chatbox AI

Install Chatbox AI, add a RunAPI custom provider, fetch a model, run the connection check, and verify a chat.

View Markdown

Connect Chatbox AI to RunAPI with a custom BYOK provider. The names below follow the current custom-provider interface; labels can vary slightly between desktop releases. Create a standard RunAPI API key in the Authentication Guide before continuing.

Install Chatbox AI

Download Chatbox AI from its official download page or the official releases, then open Settings.

Add the RunAPI provider

  1. Open AI Provider / Model Settings and select Add Provider.
  2. Name it RunAPI and choose OpenAI API Compatible as the API Mode.
  3. Enter your RunAPI API key.
  4. Set API Host to https://runapi.ai/v1 and API Path to /chat/completions.
  5. Confirm the URL preview ends in https://runapi.ai/v1/chat/completions, then save.

Keep the Host and Path separate. Fetch builds model discovery from the Host with /models, while chat uses the Host plus the configured API Path.

Add and check a model

Select Fetch to load the RunAPI model inventory. If discovery is unavailable, select New and enter an exact identifier from the Model Catalog. Chatbox AI needs at least one saved model before Check can run.

Select Check, choose the saved model, and run the connection test. Treat a successful Text Request as the baseline signal. Vision Request and Tool Use Request depend on the selected model’s capabilities and can fail independently without invalidating basic text chat.

Verify the connection

Start a new chat, choose the saved RunAPI model in the model selector, and send Reply with the word connected. A normal response confirms the custom provider is active.

Troubleshoot

  • The URL preview is wrong: use Host https://runapi.ai/v1 and Path /chat/completions; do not put the full operation URL into both fields.
  • Check is disabled or has no model: use Fetch or New to add at least one model first.
  • Text Request fails with an authentication error: update the provider with a valid standard key from the Authentication Guide.
  • Text works but Vision or Tool Use fails: choose a model that declares the required capability in the Model Catalog; these checks are not universal provider requirements.

For the shared protocol behavior, use the LLM API Quickstart. For request and response fields, use the Chat Completions API Reference.