Skip to content

Use n8n with the Gateway

n8n is an open-source workflow automation platform. Connect it to the Gateway to use Claude and other models in your automations, with your data processed in your chosen region.

Prerequisites

Connect the Gateway as an OpenAI-compatible provider

n8n's AI nodes use OpenAI-compatible APIs. The Gateway's endpoint is a drop-in replacement.

Create the credential

  1. In the left sidebar, open Overview → Credentials, then click Add credential
  2. Search for OpenAI and select it
  3. Fill in:
Field Value
API Key your LLM API key
Base URL https://llm.[your-region].amazee.ai/v1
  1. Click Save. n8n tests the connection and should show Connection successful.

You can also create the credential from inside the model node: pick Create new credential in its credential dropdown.

Use it in a workflow

When you add an AI node (e.g. AI Agent, Basic LLM Chain, Text Classifier):

  1. Click the Chat Model connector and choose OpenAI Chat Model
  2. Choose the credential you just created
  3. Pick a model from the list, for example claude-4-5-sonnet. The list comes from your Gateway, so it also shows speech and embedding models. Choose a chat model.

See Available Models for the full list.

Use Responses API toggle

Newer versions of the OpenAI Chat Model node have a Use Responses API option, and it's on by default. With it on, n8n calls /v1/responses instead of /v1/chat/completions. We have tested the Gateway with it off. If requests fail with it on, turn it off.

Connect pgvector (optional)

For RAG workflows, the Gateway provides a managed pgvector database. To get pgvector credentials for your workspace, contact ai.support@amazee.io.

In n8n, use the Postgres node or Vector Store node with:

Field Value
Host vectordb.[your-region].amazee.ai
Port 5432
Database your database name
Username your username
Password your password
SSL required

Embeddings

To generate embeddings in n8n (e.g. for a vector store workflow), use the Embeddings OpenAI node with your Gateway credential. Set the model to titan-embed-text-v2:0.

Warning

The Embeddings OpenAI node doesn't work with the Gateway right now. It runs without errors but stores unusable vectors. We're fixing this on the Gateway side.

Troubleshooting

"Model not found" or 403 team_model_access_denied
That model ID doesn't exist or isn't enabled for your key. Pick a model from the node's dropdown, which comes from your Gateway. You can also check the list with curl https://llm.[your-region].amazee.ai/v1/models -H "Authorization: Bearer YOUR_API_KEY".
Credential test fails
Make sure the Base URL ends in /v1. The test call hits /v1/models — verify that URL returns a valid response with your key.
Requests timing out on long automations
Increase the timeout in your n8n node settings. Long-running tasks (document processing, multi-step chains) may need 120–300 seconds.