Run the EveryAPI gateway inside Cherry Studio
Add EveryAPI as an OpenAI-compatible provider in Cherry Studio, then select models available to your account from the client’s model settings.
One base URL line points Cherry Studio at EveryAPI
Default global Base URL / China-first Base URL: https://api.everyapi.ai/v1 / https://api-cn.everyapi.ai/v1
# 1. Cherry Studio → Settings (gear) → Model Services# 2. Find the "OpenAI" entry (built in, no need to add one) API Key sk-everyapi-...API Host https://api.everyapi.ai/v1Status Enabled # 3. Use "Add" or "Manage Models" to check the models you want in the picker# Recommended: claude-sonnet-5 / gpt-5.5 / gemini-3.1-pro-low / MiniMax-M3From zero to your first call
- Install Cherry Studio
Download the Mac / Win / Linux installer from the official site cherry-ai.com. Open source, with data stored locally.
cherry-ai.com
- Go to Settings → Model Services
Click the gear in Cherry Studio's bottom-left → Model Services, and find the OpenAI entry.
- Change two fields
Set API Key to sk-everyapi-... and API Host to https://api.everyapi.ai/v1. Enable the provider.
- Check models
In "Manage Models" on the right of the same page, check the few models you use often. Then switch between them in the top picker when starting a new conversation in the main view.
Cherry Studio × EveryAPI
Cherry Studio ships many providers — why route through EveryAPI instead of configuring them separately?
Configuring each provider directly means applying for separate keys, dealing with each vendor's billing, and maintaining a separate integration for each. Going through EveryAPI means one key, the live model catalog, and unified billing.
To use Claude / Gemini models, do I need to create an Anthropic / Google provider?
Use the EveryAPI OpenAI-compatible provider for models supported through that protocol. Model capabilities can differ, so test the features you need with the selected model.
Are Cherry Studio's local knowledge base / agent features affected?
Not affected. Cherry Studio's local data / knowledge base / agent logic all run in the client; EveryAPI only proxies inference requests.