Amazon KeyAPI Data Workflow

Problem

Amazon data requests often get stuck between natural language and REST execution: stakeholders want category best-sellers, seller comparison, review and offer evidence, while engineers must locate endpoints, validate parameters, handle pagination, and interpret response envelopes. keyapi-amazon formalizes this translation step by inferring the goal, entity, scope, metric, pagination depth, and output, then resolving the current KeyAPI endpoint contract before making live calls.

How It Works

The skill compresses parameter-heavy APIs into a small decision model: Goal, Entity, Scope, Metric or sort, Pagination depth, and Output format. Key steps include:
- Scenario routing: choose the closest path among product discovery, category best-sellers, seller profiles, review and offer analysis, influencer storefronts, and ASIN/GTIN conversion.
- Docs-first execution: treat docs.keyapi.ai/llms.txt and endpoint pages as the source of truth for method, /v1/... path, parameters, enums, pagination, request body, and response shape.
- Controlled calls: prefer scripts/keyapi-api.mjs for live REST requests, checking HTTP status and the KeyAPI envelope where code = 0 indicates success.
- Auditable output: separate observed API facts from inference, then return a concise answer, table, raw JSON, or structured report.

Boundaries

This skill fits Amazon search, ranking, comparison, monitoring, and reporting through KeyAPI, not web scraping or undocumented endpoints. Broad reports require confirming sections and maximum scope before multi-endpoint execution. Missing KEYAPI_TOKEN means local setup must happen first. When several endpoints could work, prefer the one with lower user input burden and stronger documented filtering.