ragalgo

RagAlgo is an MCP Server that provides evaluated, scored financial context for AI agents. Instead of raw price data, it serves daily-close oriented, scored context and zones to help agents reason about market conditions after the daily close. The server exposes a set of tools for pulling news with sentiment scores, stock/crypto charts, and market snapshots. Typical usage involves connecting your MCP client to RagAlgo and requesting scored context or summaries that incorporate AI-friendly signals. You can operate RagAlgo in cloud mode (via a remote service URL) or locally (via Node.js with npx ragalgo-mcp-server --stdio) depending on your deployment needs. The core tools include get_news_scored, get_news, get_chart_stock, get_chart_coin, get_snapshots, get_financials, and search_tags, which together let you assemble a concise, decision-ready view of daily market context for AI agents.

Prerequisites:

• Node.js (recommended LTS) and npm installed on your machine
• Basic familiarity with running Node-based MCP servers

Install steps:

1. Install Node.js from https://nodejs.org/ if you don’t have it already.
2. Install the RagAlgo MCP server package (locally or via npx).
   • Local installation (optional): npm install -g ragalgo-mcp-server
   • Quick start with npx (no global install): npx -y ragalgo-mcp-server --stdio
3. Set up your API key (required for cloud/local operation):
   • Obtain an API key from RagAlgo (Dashboard or onboarding)
   • Configure environment variable RAGALGO_API_KEY with your key
4. Run the server:
   • Using npx (cloud/local): npx -y ragalgo-mcp-server --stdio
   • If you installed locally, you can run the server script directly depending on your setup
5. Verify connectivity by issuing a sample MCP request to RagAlgo (see usage in the README/documentation for endpoints and prompts).

Note: If you prefer a cloud configuration, you can point your MCP client to the RagAlgo cloud endpoint and supply the RAGALGO_API_KEY in the env.

Tips and common gotchas:

• Always provide your RAGALGO_API_KEY in the environment; without it, most RagAlgo features will be disabled or return errors.
• In cloud mode, RagAlgo serves as a remote MCP server via a URL and may not require local Node.js installation.
• The core tools prioritize scored content (e.g., get_news_scored) for AI decision-making; use get_news only if you specifically need raw data.
• When running locally, you can wrap RagAlgo in your own orchestration or proxy in front of the MCP client if you need authentication or routing logic.
• Ensure your Node.js version is compatible with the ragalgo-mcp-server package and the --stdio interface used in MCP clients.