# I wired a free no-key deals API into my agent: 7 MCP tools, verified discounts > Source: > Published: 2026-10-02 04:02:28+00:00 I built fazy.com, a price-checked deals site. Last week I needed deal data inside an agent without scraping, so I used our own API. It is free, no key, 60 req/min, and read-only. Here is the whole integration. The API: [https://fazy.com/developers](https://fazy.com/developers) The MCP server: [https://fazy.com/mcp](https://fazy.com/mcp) (7 tools: search_deals, get_deal, get_price_history, similar_deals, list_categories, catalog_slice, post_brief) Why it is different from scraping: every deal carries a verified discount_pct plus the current and reference price, and get_price_history returns the recorded price changes with dates. So the agent can answer "is this actually a good deal" instead of trusting the store's "was" price. We ran our own catalogue against those "was" prices: 83% of them were never the price while we watched (methodology: [https://fazy.com/blog/was-prices-we-never-saw](https://fazy.com/blog/was-prices-we-never-saw)). That is the dataset your agent gets to query. Minimal example, curl, no key: curl "[https://fazy.com/api/v1/deals/search?q=robot+vacuum](https://fazy.com/api/v1/deals/search?q=robot+vacuum)" Response includes discount_pct, price, reference_price, merchant, in_stock, and a url field per deal. For MCP clients, point at [https://fazy.com/mcp](https://fazy.com/mcp) and the tools show up. Repo: [https://github.com/FazyFer/fazy-mcp](https://github.com/FazyFer/fazy-mcp) I am the founder, so obviously biased, but the API is free and the data is the interesting part. Happy to answer questions about the price-tracking methodology.