ToolRouter

One MCP. 414 tools.

ToolRouter is a hosted MCP gateway: one connection gives your agent 414 specialist tools for images, video, scraping, leads, research, SEO and live data, with authentication, billing, rate limits and provider routing handled server-side, and an account auto-provisioned on first call.

Add to ClaudeUsing ChatGPT? Setup guide →

How do I install ToolRouter in one command?

claude mcp add toolrouter -- npx -y toolrouter-mcp

One command. Auto-provisioned account on first use. No key before connecting, no separate signup. Full technical docs at /docs.

What does a ToolRouter tool call look like?

Request

bash
curl -X POST https://api.toolrouter.com/v1/tools/call \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "web-search",
    "skill": "search",
    "input": { "query": "model context protocol spec", "num": 3, "provider": "serper" }
  }'

The input object is the skill's own JSON Schema: query is required, num, country, language, page, location, domains, timeRange and provider are optional, and provider accepts serper, brave, exa, dataforseo, tavily or tinyfish.

Response

Field names are the real ones; the values are illustrative.

json
{
  "id": "call_abc123",
  "status": "success",
  "output": {
    "provider": "serper",
    "provider_selection": "Serper is the default Google search provider.",
    "request_id": "req_...",
    "organic": [
      { "title": "Specification - Model Context Protocol", "link": "https://modelcontextprotocol.io/specification", "snippet": "..." }
    ]
  },
  "usage": {
    "cost": 0.0105,
    "currency": "USD",
    "credits_remaining": 9.99,
    "raw_cost": 0.01,
    "markup": 1.05
  },
  "meta": { "tool": "web-search", "skill": "search", "latency_ms": 1234 }
}

Every skill answers in that envelope: output is the skill's own schema, usage is what the call cost and what is left, meta names the tool, the skill and the latency. A failed call returns "status": "error" with an error code and message inside output, and is not billed — usage.cost is zero.

Common questions from developers

How do I add ToolRouter to Claude?

Paste the ToolRouter MCP URL into Claude Settings → Connectors, or run claude mcp add toolrouter -- npx -y toolrouter-mcp. Step by step in /docs/connect-claude.

What is ToolRouter, and how is it different from direct API integrations?

ToolRouter is a hosted gateway that sits between your AI and the tools it needs. Instead of building or managing individual API integrations — one for images, one for leads, one for search — you connect once and get all of them. ToolRouter handles authentication, billing, rate limits, and provider routing automatically.

How do accounts and API keys work?

Connector, MCP and CLI setups auto-provision an account on first use, so you do not need a key before connecting. Claim the account when you want the dashboard, credits or a named REST key — see /docs/connect-claude.

What happens when a tool call fails?

A failed paid call is never billed: the response carries a non-success status, an error message, and a usage cost of zero. Free tools never charge either way. Provider outages are routed around where a second provider covers the same skill.

Does ToolRouter work with Claude Code and other developer tools?

Yes. Claude Code, Cursor, Windsurf, VS Code Copilot, Codex CLI, and Gemini CLI all support ToolRouter via the Model Context Protocol (MCP). Add it once with a single config line and it works across every session and project automatically.

Can I bring my own keys?

Yes. If you already have an account with a provider (fal.ai, ElevenLabs, OpenAI, Perplexity, etc.) you can use your own key and skip the ToolRouter service cost entirely. Your keys are never stored — used once and discarded.

Is there a REST API?

Yes. Every tool is callable over REST at api.toolrouter.com. Create a named API key from the dashboard, send the tool + skill + input, and you get a JSON response. MCP is the fast path for agent setups; REST is there when you need direct control.

How does billing work for programmatic usage?

A paid call costs the provider's price (image generation, scraping, etc.), with a $0.01 minimum per paid call, paid from pay-as-you-go credits. Free tools stay free. The dashboard shows a per-call breakdown.

Where are the full technical docs?

See /docs for SDK reference, MCP setup across every client, REST endpoints, auth, and the model-selection hierarchy.

Ship your agent with real tools.

Add to ClaudeUsing ChatGPT? Setup guide →