Historic-Data-MCP

# Historic Price MCP Server An MCP server that looks up historic price trends for user-specified products or categories using Groq-powered URL suggestions and page scraping, then returns a structured time series and a concise trend summary. This project follows a standard MCP server structure. ## Features - URL discovery via Groq chat completions (OpenAI-compatible API) - Optional page scraping to extract price and date mentions - Heuristic parsing of prices and dates from snippets and HTML - Aggregation into a deduplicated time series and trend summary ## Tools - **price_trend_search** - Inputs: - `query` (string, required): Product/category query (e.g., "concert ticket for Shawn Mendes") - `maxResults` (number, optional, default 20): Max web results to examine (capped internally) - Output: JSON with `series` (date/price points), `summary` (direction, change, currency), and `sources`. - **scrape_price_from_url** - Inputs: - `url` (string, required): URL to scrape for price mentions and dates - Output: JSON with any extracted `series` and `meta`. ## Configuration ### Groq API Key 1. Obtain a Groq API key and set it as `GROQ_API_KEY` (see env example). 2. Optional: Set `GROQ_MODEL` (default: `llama-3.1-8b-instant`, with automatic fallback to `openai/gpt-oss-20b`) and `USER_AGENT`. Create a local `.env` (or export env vars) based on `env.example`. ### NPX (Claude Desktop / VS Code) Add this to your configuration: ```json { "mcpServers": { "historic-price": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-historic-price"], "env": { "GROQ_API_KEY": "YOUR_API_KEY_HERE", "GROQ_MODEL": "llama-3.1-8b-instant" } } } } ``` ### Docker Build and run: ```bash docker build -t mcp/historic-price:latest . docker run -i --rm \ -e GROQ_API_KEY="$GROQ_API_KEY" \ -e GROQ_MODEL="llama-3.1-8b-instant" \ mcp/historic-price:latest ``` Configure your client to run the container command as shown above. ## Development ```bash pnpm i # or npm i / bun i pnpm dev pnpm build && pnpm start ``` ## License MIT ## Reference This implementation uses Groq-assisted URL discovery and scraping to produce price history series.