Pangolinfo Amazon Data MCP

Setup takes three steps on the landing page. First, log into the Pangolinfo console and create a permanent API key under Account Center; the page warns that short-term keys expire and silently break AI workflows. Second, add one entry to your AI client's MCP configuration with the server URL, mcp.pangolinfo.com/mcp, and the key sent as a Bearer token. The transport is Streamable HTTP and nothing is installed locally. Third, restart the client, check that the tools appear in its MCP panel and ask the AI to run pangolinfo_capabilities, a free self-inspection tool that lists every live tool and its schema.

From there you work by prompting. The site offers six preset prompts, including a single-ASIN audit, a low-competition keyword scan, a trademark check that chains WIPO and PACER, and a map of a competitor seller's catalog. Each call draws credits at the same rates as the REST API.

For clients that can only launch local servers, Pangolinfo publishes an MIT-licensed stdio bridge on GitHub that forwards calls to the hosted server and runs with npx on Node.js 18 or newer.

• •Amazon core data - Search results, full product detail including title, price, bullets, A+ content and variants, paginated reviews with verified flags, seller storefronts, and a scrape_url tool for non-standard pages.
• •Category and niche analysis - Eight tools to walk the category tree, search categories, filter ASINs by price, rating or BSR, surface niches and keywords with competition scores, and pull Best Sellers and New Releases.
• •Search and AI SERP data - AI Overview and AI Mode results with their answer-card sources, keyword trend time series and Amazon Alexa search data, for spotting demand and seeing which sources AI answers cite.
• •IP and compliance checks - WIPO global trademark search for infringement screening and PACER US patent-litigation lookup with case timelines, intended for clearing a brand before listing.
• •Local maps data - Points-of-interest data from maps, used for assessing a brand's physical retail footprint.
• •Shared credits with the API - One API key and one credit balance cover the MCP, the REST API and Pangolinfo's skills, so you can move between them without a separate subscription.

The site recommends the MCP for people who do product research, product picking or one-off lookups inside an AI tool such as Claude or Cursor, who want the AI to chain tools for them, or who do not write code. Amazon sellers can use it to audit a competitor ASIN, scan for low-competition niches or check trademarks and litigation before launching a brand. Analysts and agencies can ask comparative questions across categories without building a scraper. Developers prototyping an agent can give it live Amazon data with one configuration entry.

For batch scraping, scheduled jobs or writing data into your own database, Pangolinfo itself points to its REST API instead.

Paid plans are minimum monthly commitments rather than caps, with tiered usage pricing beyond them. Credits per request vary by service, from 1 per page for product data to 5 per page for reviews, so how many calls a plan covers depends on what you ask for.