AI Agents & MCP

Connect any AI agent to all 121 ultraquick.tools tools through a single endpoint. No API key, no authentication, free, CORS-enabled. Pick the integration that fits your stack.

🔌 MCP server (recommended)

ultraquick.tools exposes every tool through the Model Context Protocol over the stateless Streamable HTTP transport. Point any MCP-compatible agent at one URL and it can list and call all 121 tools — calculators, converters, and generators — with no setup beyond the URL.

Open /mcp

Claude Desktop

Add this to your claude_desktop_config.json (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "ultraquick": {
      "url": "https://ultraquick.tools/mcp"
    }
  }
}

Cursor

Settings → MCP → Add new MCP server → Type HTTP, URL below:

https://ultraquick.tools/mcp

Claude Code / Cline / VS Code (MCP)

Add the server from the command line or settings — a streamable-HTTP server needs only the URL:

claude mcp add --transport http ultraquick https://ultraquick.tools/mcp

Any MCP client (JSON config)

{
  "mcpServers": {
    "ultraquick": { "url": "https://ultraquick.tools/mcp" }
  }
}

Try it raw (JSON-RPC)

The MCP server speaks JSON-RPC 2.0. List every tool, then call one:

curl -X POST https://ultraquick.tools/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"demo","version":"1.0.0"}}}'

curl -X POST https://ultraquick.tools/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'

curl -X POST https://ultraquick.tools/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"calculate-bmi","arguments":{"unit":"metric","height":170,"weight":70}}}'

⚡ OpenAI function-calling

Prefer native function-calling? Download the ready-made definitions (OpenAI tools format) and pass them straight to your model. Each entry's description includes the endpoint, so the model knows where to POST.

Download functions.json

import json, urllib.request

defs = json.load(open("functions.json"))  # 119 OpenAI tool definitions
# pass `defs` as `tools=defs` to your model, then POST the chosen args to the endpoint
# in the tool description.

🗺️ OpenAPI, llms.txt & the full manifest

OpenAPI 3.1 — every endpoint, parameter, and response schema.
llms.txt — concise index for LLMs.
llms-full.txt — detailed per-tool reference for LLMs.
ai.json manifest — the whole registry in one JSON file (tools, endpoints, schemas).
sitemap.xml

https://ultraquick.tools/ai.json

How it works

Each ultraquick.tools tool has a pure compute function shared by its web UI and its JSON API. The MCP server imports those same functions, so results are byte-identical whether you use the website, the REST API, or MCP. The server is stateless and state-free — no sessions, no rate limits, no data stored — which keeps it fast and free on Cloudflare's edge.

No authentication is ever required. Requests are idempotent and side-effect-free, so they're safe for agents to call freely. CORS is permissive (Access-Control-Allow-Origin: *), so browser-based agents work too.