Skip to main content

Cherry Studio

Cherry Studio connects through an OpenAI compatible provider with the /v1 Base URL.

Before configuring​

  • This page uses placeholders only. Copy the real Key from PopGo API keys and the model from models and pricing.
  • One-click setup is for software that is already installed and has registered its protocol handler. Allow the browser to open the external app when prompted.
  • For manual setup, save one minimal Provider first. Enable streaming, tools, or multi-model routing only after the first request succeeds.

Official source and verification​

This page was verified on 2026-07-13 against the official source for Cherry Studio 1.9.12. Official source. Verification covers Cherry Studio custom provider setup.

One-click setup​

Sign in to the PopGo API keys page and open the action menu on the target Key row. Public documentation never creates or displays a link containing a Key. Open the Chat submenu, then choose Cherry Studio. After import, confirm the provider is OpenAI compatible and select a currently available model.

Manual setup​

Choose an OpenAI compatible provider and enter the fields below.

FieldValue
Provider typeOpenAI compatible
Base URLhttps://api.popgo.site/v1
API KeyYOUR_API_KEY
ModelYOUR_MODEL

Text-only setup flow​

  1. Install or update Cherry Studio to the current version, then open “Settings → Model services”.
  2. Create a custom Provider and choose OpenAI compatible. Do not modify the client’s built-in official Provider.
  3. Enter https://api.popgo.site/v1 as Base URL, YOUR_API_KEY as API Key, and YOUR_MODEL from models and pricing as Model.
  4. Save, create a new conversation, and send a short message first. Enable image, video, tools, or long context only after it succeeds.
  5. Create separate Providers for different protocols or model families so Claude, Gemini, and OpenAI compatible addresses are not mixed.

Common setup symptoms​

  • An empty model list does not necessarily mean the connection failed. Many clients do not fetch third-party models; enter YOUR_MODEL manually.
  • If saved settings still call the default service, switch to the new Provider, close the old conversation, and restart the client.
  • If a new capability is missing in an older client, update it and follow the fields shown by the current UI.

Success check​

Send a short message from the model test or a new conversation. Returned content without an error confirms success.

Common errors and recovery​

  • If connection testing fails, check that /v1 is not duplicated at the end of the Base URL.
  • For a 401, paste YOUR_API_KEY again and remove surrounding whitespace.
  • For a 404, verify the Base URL level and make sure the client is not appending /v1 twice.