🦞OpenClaw provider setup

Tokaroo + OpenClaw

One API key. Real-time AI routing across 199+ AI models from Anthropic, Google, Groq, and OpenAI - chat, images, speech, and video - optimized for cost, latency, and automatic failover. Connect in under 60 seconds.

Install the published Tokaroo provider plugin in one command, or keep the manual OpenAI-compatible provider block as a fallback.

Get your API key View docs

Recommended setup: plugin install

This gives you the native OpenClaw plugin flow and keeps Tokaroo discoverable in the OpenClaw ecosystem.

Create your API key

Get API key ->

Install the Tokaroo plugin

This is the recommended native OpenClaw setup path.

openclaw plugins install clawhub:tokaroo-openclaw-provider

Set your API key

Add it to your environment or OpenClaw's .env file.

# ~/.openclaw/.env  (or export in your shell)
TOKAROO_API_KEY=tok_...

Restart OpenClaw

Select tokaroo/auto and every agent call routes through Tokaroo.

openclaw restart

Fallback: manual OpenAI-compatible config

If you do not want the plugin, Tokaroo still works as a manual OpenAI-compatible provider in OpenClaw.

Add Tokaroo to your openclaw.json

Open ~/.openclaw/openclaw.json and add the tokaroo provider block. OpenClaw uses JSON5 - comments are allowed.

// ~/.openclaw/openclaw.json
{
  models: {
    mode: "merge",
    providers: {
      tokaroo: {
        baseUrl: "https://api.tokaroo.com/v1",
        apiKey: "${TOKAROO_API_KEY}",
        api: "openai-completions",
        models: [
          { id: "auto", name: "Tokaroo Auto - best value, automatic routing" },
          { id: "fast", name: "Tokaroo Fast - speed-optimized, low latency"  },
          { id: "max",  name: "Tokaroo Max  - highest capability"             },
        ]
      }
    }
  },
  agents: {
    defaults: { model: "tokaroo/auto" }
  }
}

Set your API key

Add it to your environment or OpenClaw's .env file.

# ~/.openclaw/.env  (or export in your shell)
TOKAROO_API_KEY=tok_...

Restart OpenClaw

That's it. Every LLM call in your agents now routes through Tokaroo.

openclaw restart

What your agents get

Every model, one key

Every major model across OpenAI, Anthropic, Google, Groq, Meta, Mistral, DeepSeek, and Cohere - all available through a single Tokaroo API key. Use tokaroo/auto and let us pick the best one.

Semantic cache

Repeated or similar questions can be served from cache when eligible, which cuts repeat latency and cost.

Centralized routing

Tokaroo keeps cache, local, BYOK, and managed routing decisions in one place so your agents do not need provider-specific branching.

Usage analytics

See exactly what your OpenClaw agents spent over time. Usage, latency, and billing are visible at tokaroo.com/dashboard.

Local + cloud hybrid

Run Ollama or vLLM locally. Tokaroo handles the routing - local for simple tasks, cloud for complex ones.

Works with every OpenClaw skill

No changes to your skills or agent configs beyond the provider block. Everything else stays the same.

With local models (Ollama / vLLM)

Already running Ollama locally? Tokaroo can route simple requests to your local models and fall back to cloud for complex tasks - automatically.

// With local Ollama models alongside Tokaroo
{
  models: {
    mode: "merge",
    providers: {
      tokaroo: {
        baseUrl: "https://api.tokaroo.com/v1",
        apiKey: "${TOKAROO_API_KEY}",
        api: "openai-completions",
        models: [
          { id: "auto", name: "Tokaroo Auto" },
          { id: "fast", name: "Tokaroo Fast" },
          { id: "max",  name: "Tokaroo Max"  },
        ]
      },
      local: {
        baseUrl: "http://localhost:11434/v1",
        apiKey: "local",
        api: "openai-completions",
        models: [
          { id: "llama3.2", name: "Llama 3.2 (local)" },
          { id: "mistral",  name: "Mistral (local)" }
        ]
      }
    }
  },
  agents: {
    defaults: { model: "tokaroo/auto" }
  }
}

Ready to connect?

Create your API key and have your agents routing through Tokaroo in under a minute.

Create API key