Sugra API MCP

What is Sugra API MCP?

Sugra API MCP is an official Model Context Protocol server that acts as a gateway connector between LLM agents and world data via the Sugra API. It runs on Python 3.11+ and works with Anthropic Claude, OpenAI ChatGPT, Google Gemini, xAI, and any MCP-enabled IDE.

How to use Sugra API MCP?

Install with pip install sugra-api-mcp, obtain a free API key from app.sugra.ai/settings/billing, then configure the server in your MCP client’s settings file (e.g., claude_desktop_config.json for Claude Desktop, ~/.claude/config.json for Claude Code, or mcp.json for Cursor, Zed, etc.). For remote access, use the hosted HTTP endpoint https://app.sugra.ai/mcp with Bearer authentication. CLI commands like sugra-api-mcp doctor, sugra-api-mcp search, and sugra-api-mcp call are also available.

Key features of Sugra API MCP

• Six MCP tools for data retrieval and catalog browsing
• Combines search and call in one step with fetch_data
• Supports response shaping: limit, fields, include_raw
• Works with multiple AI clients (Claude, ChatGPT, Gemini, xAI)
• Offers both stdio and Streamable HTTP transport
• Includes CLI helpers: doctor, search, describe, call

Use cases of Sugra API MCP

• Ask an AI agent to search Sugra endpoints for NASDAQ futures
• Describe an operation like cot_financial to understand its inputs
• Call quotes_symbol_price to get real-time stock data for a symbol
• List available Sugra toolsets to discover data groups
• Use fetch_data to query world data with natural language

FAQ from Sugra API MCP

How do I install Sugra API MCP?

Install via pip: pip install sugra-api-mcp. Requires Python 3.11 or higher.

How do I get an API key?

Get a free API key at app.sugra.ai/settings/billing. The free tier allows 50 requests per day.

What transport modes are supported?

Two modes: stdio (local process, default) and Streamable HTTP (remote). Use --transport streamable-http --port 8001 for HTTP mode.

Which AI clients are supported?

Anthropic Claude (Desktop, Code, web), OpenAI ChatGPT, Google Gemini, xAI, and IDEs such as VS Code, Cursor, Zed, Cline, Continue.dev, and Windsurf.

What happens if I hit the rate limit?

You receive a 429 response. The headers include X-RateLimit-Reset with the UTC timestamp when the daily counter resets (midnight UTC). Upgrade your plan or wait for the reset.