Using Downsizing with n8n

Prerequisites

A running n8n instance (self-hosted or n8n Cloud) and a Downsizing account that can make requests (pay-per-use or a connected subscription both work). Grab your Inference ID and API key from the Inference API page.

Manual Setup

n8n's AI nodes come from its LangChain integration. The Anthropic Chat Modelnode/credential only exposes an API key field (plus an optional custom-header toggle) — there is no base URL override, so it can't be pointed at a proxy. The OpenAI Chat Modelnode's credential does support a custom endpoint via a Base URL field, so use that node to route through Downsizing instead:

1. In your workflow, add an OpenAI Chat Model node (under AI → Language Models).
2. Create a new credential of type OpenAi API. In the credential form, set:
   • API Key: your Downsizing API key (starts with dws_).
   • Base URL: your Downsizing OpenAI-compatible endpoint.

http://localhost:3001/YOUR_INFERENCE_ID/openai

Replace YOUR_INFERENCE_ID with your workspace UUID from the dashboard. On the OpenAI Chat Model node itself, if a Use Responses API toggle is available in your n8n version, enable it — Downsizing prefers the Responses API over Chat Completions.

Verify

Run the workflow (or use the node's test-step action) and confirm you get a completion back. Then check your savings dashboard to see the request appear in your usage metrics.