Skip to Content
Raydo

Models and providers

Set up Raydo Credits, use your own API key, or connect Ollama or ComfyUI.

A provider is the service that runs a model. Open Models in the full workspace sidebar for managed and BYOK connections. Go to Settings → Local for Ollama and ComfyUI. App connections and MCP supply separate tools or data for a task.

Use Raydo Credits

  1. Go to Settings → Cloud and select Login via Browser. Complete sign-in and return to Raydo.
  2. Check the connection and credit balance. In a new Chat, select an available model from the managed service and send a short test prompt.

The managed connection uses your Raydo account. If you are entering a WeToken base URL for an OpenAI-compatible model API client or SDK, use https://wetoken.one/v1.

Bring your own API key

  1. From Models, choose Add Provider and select the service you have an account with. Use the listed endpoint for an official provider.
  2. Enter the key in API Key. For a custom compatible service, fill in API URL and choose the matching API type.
  3. Add or select a model. Use Fetch model list when available, then Test Connection for the selected model and save.
  4. Select that model in a new chat and send a short prompt with no private information.

OpenRouter processes requests and bills your OpenRouter account separately from Raydo Credits. To replace a legacy connection, verify the new model first and then update any default, role, or workflow bindings that use the old model.

Connect Ollama or ComfyUI

Under Settings → Local, enter the address of a running service on your computer or private network. On the same computer, typical addresses are http://localhost:11434 for Ollama and http://127.0.0.1:8188 for ComfyUI. Enter the service address by itself, without an API path. Choose Test, check the returned models or checkpoints, then choose Save.

  • For Ollama, select one of the saved models for a text task. Open Models takes you to the model page, where you can also change the default.
  • For ComfyUI, select the service and a checkpoint in the image workspace and run a small generation.

If the test fails, check that the service is running and the address is reachable from Raydo. If it connects with an empty list, install an Ollama model or add a ComfyUI checkpoint and test again.

On this page

Models and providers