Skip to main content
Connect OpenClaw to Eden AI and route your agent through 500+ models with automatic fallbacks and a single API key.

Overview

OpenClaw is a terminal-based AI coding agent. By pointing it at Eden AI, you get:
  • 500+ models: Access OpenAI, Anthropic, Google, and more through a single API key
  • One key, many providers: Switch models by changing a single config value — no per-provider keys
  • Automatic fallbacks: Requests reroute to healthy providers if one goes down
Prefer a video walkthrough? Follow along on YouTube.

Prerequisites

Setup

1. Add Eden AI as a custom provider

OpenClaw stores its config at ~/.openclaw/openclaw.json (on Windows: %USERPROFILE%\.openclaw\openclaw.json). Open it in your editor of choice and add the Eden AI provider with a function-calling model:
Use any provider/model from the Eden AI catalog. Pick one that supports function calling, since OpenClaw sends tools with every request.
A few notes on the config above:
  • api must be openai-completions. Eden AI exposes an OpenAI-compatible endpoint, and this tells OpenClaw which wire protocol to use. If you omit it, the model fails to load with no "api" specified.
  • agents.defaults.model.primary makes the Eden AI model OpenClaw’s default. It must be edenai/<model-id> — the provider key (edenai) plus the model’s id (e.g. edenai/anthropic/claude-sonnet-4-5).
  • name is just the display label shown in OpenClaw’s model picker. If you switch id to another model, rename this field too so the picker doesn’t show a stale label.
  • cost values are placeholders. Billing is handled on the Eden AI side through your API key, so OpenClaw’s local cost tracking isn’t used here. Check per-request cost in the Eden AI dashboard.
  • supportsTools: true enables tool/function calling through Eden AI. Set it to false only if you want to restrict this provider to plain chat completions.

2. Configure OpenClaw interactively

Run the built-in wizard:
In the Model and auth step, select Eden AI — Claude Sonnet 4.5 — the model you defined in Step 1 (it’s already your default from the agents.defaults.model.primary setting). Accept the defaults for the remaining steps (Workspace, Gateway, Channels, etc.) unless you have specific local-networking or channel needs.
The wizard’s Gateway step configures OpenClaw’s local server (port, bind address, auth mode, tailscale exposure) — it does not ask for the Eden AI API URL. That’s already set as models.providers.edenai.baseUrl in Step 1.

3. Start OpenClaw

Every LLM call now flows through Eden AI, with built-in fallbacks, unified billing, and per-request cost visibility.

Using specific models

Target any model by setting the id to a provider/model string:
Common choices: Browse the full catalog at app.edenai.run/models.

Troubleshooting

401 Unauthorized

Verify your Eden AI key — grab a fresh one at app.edenai.runAPI Keys. Check for leading/trailing spaces.

model not found

Ensure the id uses the provider/model format (e.g. anthropic/claude-sonnet-4-5). No bare model names.

no "api" specified

The provider is missing its api field. Set "api": "openai-completions" at the provider level (see Step 1) — Eden AI uses the OpenAI-compatible protocol.

Connection issues

Confirm the base URL is exactly https://api.edenai.run/v3 and check Eden AI status at app-edenai.instatus.com.

Next Steps