Skip to content

Setting up Cherry Studio

This page is for the Cherry Studio mobile app on iOS and Android. The app is in development, so button names can move. Trust the request-address preview on the setup form.

The desktop app is a separate product. Pairing can import connection settings. It does not import chat history.

One Nexus key is enough. Each protocol is its own connection entry in the app, not a second Nexus account. The shared order is on Client setup.

Values

Entry What you select Base URL to try first Preview must show
Chat OpenAI Chat https://api.nexus-hub.tech https://api.nexus-hub.tech/v1/chat/completions
Responses OpenAI Responses https://api.nexus-hub.tech https://api.nexus-hub.tech/v1/responses
Messages Anthropic Messages https://api.nexus-hub.tech https://api.nexus-hub.tech/v1/messages
Gemini Gemini Generate Content https://api.nexus-hub.tech a path with /v1beta/models/ and generateContent

If the preview is missing /v1 or shows the operation path twice, change only the base until the preview matches the table. For Chat and Responses, that usually means setting the base to https://api.nexus-hub.tech/v1. For Messages, keep the base without /v1. For Gemini, the preview must contain /v1beta, not only /v1.

The form rejects a full operation path, a query string, and a key inside the URL. Do not add a trailing #. That mark turns off version insertion, and Nexus needs the normal version path.

Steps

  1. Create a Nexus API key in the Telegram bot.
  2. In the app, open model-service settings and add a custom entry named Nexus.
  3. Select OpenAI Chat. Set the base URL from the table and read the request-address preview before you save.
  4. Paste the key with no Bearer prefix and no spaces.
  5. Sync models, or type an exact id from GET /v1/models for this key.
  6. Enable the entry and send a short message.
  7. Add a second entry for Anthropic Messages. The preview must end in /v1/messages.
  8. Add a third entry for Gemini only if you need that format. The preview must contain /v1beta/models/ and generateContent.
  9. Add a Responses entry only when you explicitly select that interface. The preview must end in /v1/responses.

Tools, attachments, and images work only when the catalog allows them for that id and route. A model id from Chat is not automatically valid on Messages or Gemini.

Check

Send one short message on the entry you just enabled. A normal reply means that entry works. Repeat the check after you add Messages or Gemini.

If it does not work

  • 401 or invalid_key: the key field has extra spaces, a Bearer prefix, or a revoked key.
  • model_not_found or model_not_allowed: the id is not in this key's GET /v1/models list.
  • unsupported_model: the id is visible, but not on this protocol. Use the entry whose preview matches the route family, or pick another id.
  • 404: the preview is not one of the four addresses in the table. Do not paste the operation path into the base field.

Send support the route, error code, time, and x-request-id. Do not send the key or a full prompt.

Official app: Cherry Studio for iOS and Android.