Description
n8n-nodes-edenai
This is an n8n community node for Eden AI — a European AI gateway that gives you access to 300+ models from 50+ providers through a single API and credential.
Features
- 300+ models from providers like OpenAI, Anthropic, Mistral, Google, and more
- Automatic fallback — define backup models if the primary fails
- Web search — enable real-time web access for supported models
- European hosting — full transparency on where models are hosted, GDPR-compliant
Installation
In your n8n instance, go to Settings → Community Nodes and install:
n8n-nodes-edenai
Credentials
You need an Eden AI API key. Get one at app.edenai.run.
Usage
This package provides three nodes:
Eden AI Chat Model
Add the Eden AI Chat Model node as a sub-node to any AI Agent or Chain in your workflow.
#### Options
| Option | Description |
|——–|————-|
| Model | Provider/model format (e.g. openai/gpt-4o, anthropic/claude-sonnet-4-5) |
| Fallback Models | Comma-separated backup models if the primary fails |
| Web Search | Enable real-time web access (model must support it) |
| Web Search Context Size | low, medium, or high — controls retrieval depth |
| Temperature | Sampling temperature (0–2) |
| Max Tokens | Maximum tokens to generate |
| Response Format | text or json_object |
Eden AI Expert Models
The Eden AI Expert Models node runs Eden AI’s non-LLM features through the Universal AI endpoint. Unlike the Chat Model and Embeddings sub-nodes, it’s a regular action node — connect it inline in a workflow.
Pick a Feature, Subfeature, and Provider (loaded live from your account), set the Input Type to Text or File (a public URL, an uploaded file ID, or binary data from a previous node), and add any extra parameters (e.g. targetlanguage, voice, documenttype) under Additional Input Fields. Toggle EU Only to route through api.eu.edenai.run and list only EU-eligible providers (GDPR / data residency).
Other Options:
| Option | Default | Purpose |
| — | — | — |
| Fallback Models | — | Comma-separated providers (up to 3) tried in order if the primary fails. |
| Provider Params | {} | Provider-specific parameters sent as provider_params; takes precedence over the normalized input fields. |
| Simplify Response | false | Return only the feature output instead of the full envelope (status, cost, provider, output). |
| Show Original Response | false | Include the raw provider response under original_response. |
| Download File Output | false | Download generated files (image generation, text-to-speech, document translation) and attach as binary data under Output Binary Field (default data). |
Asynchronous features — speech-to-text, multi-page OCR (ocrasync, ocrtables_async), and video generation — are supported and labelled (async) in the Subfeature list. By default the node launches the job and polls until it finishes. Control this under Options:
| Option | Default | Purpose |
| — | — | — |
| Wait for Completion | true | Poll until the job finishes and return its result. Turn off to return the job handle (public_id) immediately and poll it yourself or receive a webhook. |
| Poll Interval (Ms) | 4000 | Delay between job status checks while waiting. |
| Max Wait Time (Ms) | 300000 | How long to poll before timing out. |
| Webhook URL | — | A URL Eden AI calls when the job completes (sent as webhook_receiver). |
This node is also exposed as an AI tool (usableAsTool), so an AI Agent can call Eden AI expert models directly.
