Ecommerce Data API

Ecommerce Data API is an Amazon data interface service for cross-border ecommerce sellers and developers, with 46 REST endpoints covering products, keywords, market data, traffic and trademarks. It is deterministic data retrieval: no generative model runs on the server, and AI appears only at the integration layer and in a single OCR endpoint.

Core capabilities

• 46 data endpoints: 12 product, 12 keyword, 14 market, 3 traffic, 4 trademark and 1 OCR, all POST, documented with OpenAPI 3.1
• Hosted remote MCP: a single /mcp address carries 45 business tools plus 4 general tools, with OAuth 2.1 + PKCE, so Claude Code, Codex, Cursor and ChatGPT can connect directly
• Official CLI and Skill: install.sh downloads a SHA-256-verified local stdio client and writes the Skill into Claude Code, Codex and Cursor
• Product research and trends: ABA weekly and monthly reports, reverse lookup of traffic keywords, BSR sales trends (table lookup against preset top-level category curves; categories without a curve return 422)
• 13 marketplaces, 24 months: all 13 marketplaces including US, UK, DE and JP are open, and date parameters reach back up to 24 months
• AI capability library: catalogs 35 third-party Skills/Agents with author, license and source noted, and does not show ratings

Where it fits Data developers who do not want to build their own scrapers but need stable field names; heavy AI users running product and competitor analysis in Claude Code or Cursor; small and mid-sized teams studying category concentration and ABA keyword lists; cross-border brands doing a first-pass trademark screen.

What sets it apart Rainforest starts at $23/month, while Keepa starts at €49/month and charges by token rate; the differentiation here sits in the Agent integration layer beyond REST, and the Chinese documentation also suits Chinese-speaking sellers better. The trade-off is that sales and search volume are third-party estimates.

Editorial verdict Documentation quality runs ahead of product maturity: every endpoint has a field table, examples, error codes and rate-limit headers, and both the no-charge-on-failure rule and the earliest-expiry-batch-first deduction are stated clearly. Yet corporate and legal details are almost blank, the licensing chain is opaque, and the product has been live for only a month or two with almost no independent reviews, so buying the smallest pack first to check data freshness is the steadier move.