---
title: "Eden AI: the independent European AI gateway"
description: "One OpenAI-compatible API to around 1,000 models from 80+ providers, with cost-based routing, spend control per team and a European endpoint."
canonical: https://www.edenai.co/
last-updated: 2026-11-02
---
# Eden AI: the independent European AI gateway

> One OpenAI-compatible API to around 1,000 models (LLMs and expert models) from 80+ providers, with cost-based routing, spend control per team and a European endpoint.

## What Eden AI does

- **AI gateway and LLM router:** one API key and one request shape for every model. Route by cost, speed or latency, declare fallbacks, cache repeated requests.
- **FinOps:** every response includes its cost in USD. Track spend by team, product or end user, and give each API key a budget that resets daily, weekly or monthly.
- **Governance:** tags on every call, model allow-lists, roles, and on the Advanced plan SSO, SCIM and audit logs.
- **Expert models:** OCR, document parsing, speech-to-text, text-to-speech, translation and vision through the same key.

## Quickstart (OpenAI SDK)

```python
from openai import OpenAI

client = OpenAI(base_url="https://api.edenai.run/v3", api_key="YOUR_EDEN_AI_API_KEY")

response = client.chat.completions.create(
    model="openai/gpt-5.5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)
```

Model IDs use the `provider/model` format. An Anthropic-compatible messages endpoint is also available.

## Pricing

- Provider prices with 0% markup.
- 5.5% platform fee, only when buying credits.
- No subscription and no API call limit.
- Advanced plan: custom terms, volume discounts, SSO, SCIM, audit logs, custom billing.

## Data and compliance

- EU endpoint: `https://api.eu.edenai.run/v3`. Requests naming a model not cleared for European processing are refused with HTTP 451 before any provider is contacted.
- Zero data retention by default; no training on customer data.
- SOC 2 and ISO 27001 certified; Data Processing Agreement available.
- French company based in Lyon, founded in 2021. Used by 300k+ developers.

## For agents

- MCP server: https://mcp.edenai.run/mcp (streamable HTTP, Bearer API key). Expert models as tools (OCR, web search, speech, translation, image, video) plus `list_models`, `upload_file` and `check_job`. Docs: https://www.edenai.co/docs/v3/expert-models/mcp-server
- Agent skills: https://www.edenai.co/.well-known/agent-skills/index.json
- Catalog of everything we offer agents: https://www.edenai.co/.well-known/ai-catalog.json
- API catalog: https://www.edenai.co/.well-known/api-catalog
- OpenAPI spec: https://api.edenai.run/v3/docs/openapi.json

## Links

- Documentation: https://www.edenai.co/docs
- Documentation index for LLMs: https://www.edenai.co/docs/llms.txt
- Model catalogue: https://www.edenai.co/models
- Pricing: https://www.edenai.co/pricing
- Get an API key: https://app.edenai.run/user/register
- Contact: https://www.edenai.co/contact

_Last updated 2026-11-02._