Amazon Blue Ocean Market Detector

Problem

Amazon product selection often fails not because keywords are missing, but because signals are scattered: category search volume, niche competition, ASIN review counts, sales velocity, new-release momentum, off-site demand, design patents, and trademark risk are usually evaluated in isolation. Manual spreadsheet work can mistake high search volume for opportunity, or treat a zero-result long-tail query as proof that demand does not exist. amazon-blue-ocean-market-detector turns this into a structured go/no-go workflow: validate demand, assess competitive structure, and filter out high-risk directions using IP and review signals.

How It Works

• Fast mode: the default path, targeting roughly 90 seconds.
• Uses filter_niches and keyword_trends to check category or keyword momentum.
• Uses search_amazon to inspect the first five organic results and identify a Moat Giant (high review count, mature badges) and a Breakout Black Horse (lower review count but strong sales or ranking signals).
• Uses list_new_releases and get_amazon_product to measure new-entry pressure and single-product detail.
• Produces a structured five-part report instead of dumping raw JSON.
• Full mode: triggered only when the user explicitly asks for a deep or complete report.
• Uses ai_search to collect off-site demand signals from Reddit, TikTok, and similar sources.
• Uses wipo_search to scan USPTO design patents and flag active patents held by large companies or law firms.
• Uses ai_search for a preliminary text-trademark screening pass.
• After budget confirmation, calls get_amazon_reviews and clusters the top negative-review pain points.
• Data discipline: the report only uses hard data returned by the MCP tools. Key numbers are traceable to calls such as search_amazon, get_amazon_product, and filter_niches; missing fields are explicitly marked as not returned by the backend instead of being filled with approximations.

Boundaries

This skill is useful for early product discovery, category entry decisions, competitor prototyping, and pre-launch risk scanning. It is not a substitute for formal legal clearance or investment advice. Fast mode avoids slower, costlier tools such as review scraping. marketplaceId must be an ISO site code such as US, UK, or DE, not an Amazon merchant ID. Terminal errors like AUTH, QUOTA, and BAD_INPUT should not be blindly retried; transient errors such as RATE_LIMIT or 9200 require lowering concurrency and retrying only the failed request.