CherryIN Docs

Call the CherryIN API from n8n using the HTTP Request node for automation and workflow use cases

Why use the HTTP Request node?

n8n includes built-in OpenAI integrations, but the available fields and behavior can differ across versions. To keep this guide stable and reusable, we use the HTTP Request node first, which is also the universal fallback recommended in n8n's docs.

This gives you two benefits:

It works across almost all n8n versions
You fully control the endpoint, request body, and model ID

Before you start

Please prepare:

A working n8n instance
Your CherryIN API key
A model ID such as anthropic/claude-sonnet-4.5

Setup Steps

Create a Bearer Auth credential

Create a generic Bearer Auth credential and paste your CherryIN API key into it.

If your UI doesn't make that obvious, this credential is simply sending:

Authorization: Bearer sk-xxxxxxxx

Create a Chat Completions request

Add an HTTP Request node and configure it like this:

Field	Value
`Method`	`POST`
`URL`	`https://open.cherryin.net/v1/chat/completions`
`Authentication`	`Bearer Auth`
`Content-Type`	`application/json`

For your first test, use this request body:

{
  "model": "anthropic/claude-sonnet-4.5",
  "messages": [
    {
      "role": "user",
      "content": "Hello, introduce CherryIN in one sentence."
    }
  ]
}

If the node returns a model response, the integration is working.

Use data from previous nodes

Once the connection works, you can inject upstream data into the request body. For example:

{
  "model": "google/gemini-2.5-flash",
  "messages": [
    {
      "role": "user",
      "content": "Summarize this text: {{$json.content}}"
    }
  ]
}

This lets you pass form input, database rows, scraped content, or any other workflow output into a CherryIN model.

How do I use GPT-5.x models?

If you want to call GPT-5.x models in n8n, switch to the Responses API:

Field	Value
`Method`	`POST`
`URL`	`https://open.cherryin.net/v1/responses`
`Authentication`	`Bearer Auth`
`Content-Type`	`application/json`

Example request body:

{
  "model": "openai/gpt-5.2-chat",
  "input": "Turn this content into 3 bullet points: {{$json.content}}"
}

Why use Responses here?

GPT-5.x models should not be forced through Chat Completions. If you're sending raw HTTP requests in n8n, the safest path is /v1/responses.

Chat Completions: https://open.cherryin.net/v1/chat/completions
Responses API: https://open.cherryin.net/v1/responses

My expression didn't evaluate

Make sure you're using n8n expression syntax such as {{$json.content}}, and that the node is sending a JSON body.

Next steps

Want the universal setup pattern first? See OpenAI-Compatible Setup Guide
Want app orchestration next? See Dify Setup Guide

n8n Setup Guide