CherryINCherryIN
New API Guides

OpenAI-Compatible Setup Guide

Use CherryIN with Cline, Dify, n8n, Chatbox, CherryStudio, and other OpenAI-compatible tools

Who is this for?

If your tool supports any of the following setup styles, you can use this guide directly:

  • OpenAI
  • OpenAI Compatible
  • Custom OpenAI
  • A manual Base URL + API Key setup

Typical examples include Cline, Dify, n8n, CherryStudio, Chatbox, and most AI clients or workflow tools that support the OpenAI protocol.

You only need 3 values

FieldWhat to enterExample
Base URLCherryIN's OpenAI-compatible endpointhttps://open.cherryin.net/v1
API KeyThe token you created in the CherryIN consolesk-xxxxxxxx
Model IDFull model name in provider/model formatanthropic/claude-sonnet-4.5

Token Group Tip

For most users, a default group token is the easiest option because it can access more models directly. If you want discounted models from a promotional group, make sure the token and model group match. See Quick Start.

What these fields are usually called in tools

Tool field nameWhat you should enter
API Provider / ProviderSelect OpenAI, OpenAI Compatible, or Custom OpenAI
Base URL / API Base URL / Endpointhttps://open.cherryin.net/v1
API Key / Bearer TokenYour CherryIN API key
Model / Model IDFor example anthropic/claude-sonnet-4.5
Organization IDUsually leave blank

Universal setup template

  1. Choose an OpenAI or OpenAI Compatible provider in your tool.
  2. Set Base URL to https://open.cherryin.net/v1.
  3. Paste your CherryIN key into API Key.
  4. Enter a full model ID such as anthropic/claude-sonnet-4.5.
  5. Save, test, or send a message to verify the setup.

Model Name Format

CherryIN model IDs must use the full provider/model format. For example, use anthropic/claude-sonnet-4.5, not just claude-sonnet-4.5.

When should you use Chat Completions vs Responses?

ScenarioRecommended API
Most chat tools, IDE assistants, and workflow nodesChat Completions
A tool lets you choose the raw API mode and you want to use a GPT-5.x modelResponses API

GPT-5.x Compatibility

For GPT-5.x models such as openai/gpt-5.2-chat, prefer the Responses API instead of Chat Completions.

Chat Completions example

curl https://open.cherryin.net/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxx" \
  -d '{
    "model": "anthropic/claude-sonnet-4.5",
    "messages": [
      {"role": "user", "content": "Hello, introduce yourself in one sentence."}
    ]
  }'

Responses API example

curl https://open.cherryin.net/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxx" \
  -d '{
    "model": "openai/gpt-5.2-chat",
    "input": "Summarize CherryIN setup in 3 bullet points."
  }'

Common issues

1. 401 Unauthorized

Usually caused by one of these:

  • The API key was copied incorrectly
  • There is an extra space or newline in the key
  • The tool isn't sending the key as a Bearer token

2. Model not found

Make sure you're using the full model ID, for example:

  • Correct: anthropic/claude-sonnet-4.5
  • Incorrect: claude-sonnet-4.5

3. The tool only says OpenAI, not OpenAI Compatible

That can still work if the tool lets you manually set a custom Base URL. If it doesn't allow a custom endpoint at all, it usually can't connect to CherryIN directly.

4. GPT-5.x request fails

Check whether the tool is hardcoded to /chat/completions. If so, switch to a setup that supports the Responses API, or test with a non-GPT-5.x model first.

Next steps

Code Tutorial

Step-by-step guide to calling the CherryIN API with code

Claude Code Setup Guide

Complete guide to using CherryIN API with Claude Code

On this page

Who is this for?You only need 3 valuesWhat these fields are usually called in toolsUniversal setup templateWhen should you use Chat Completions vs Responses?Chat Completions exampleResponses API exampleCommon issues1. 401 Unauthorized2. Model not found3. The tool only says OpenAI, not OpenAI Compatible4. GPT-5.x request failsNext steps