First public release of KDP Niche Scorer MCP — an AI-native Amazon KDP niche research server.
What's included
- 3 MCP tools: `score_niche`, `get_top_competitors`, `estimate_sales_from_bsr`
- 0–100 opinionated scoring ported verbatim from the KDP Research Chrome extension: 35 demand + 30 competition + 15 pricing + 10 indie/recency, minus up to 25 for top-3 dominance
- Kindle 70% / paperback 60% royalty modes with calibrated BSR-to-sales lookup (Nov 2025)
- Defensive RapidAPI client with 24h TTL cache, 5 marketplace support (US, UK, DE, CA, AU)
- 22 unit tests + MCP protocol smoke test — all green
- FastMCP 2.14, Python 3.11+, Streamable HTTP transport
Install (hosted on MCPize)
```bash
claude mcp add --transport http kdp-niche-scorer https://kdp-niche-scorer-mcp.mcpize.run/mcp
```
Marketplace: https://mcpize.com/mcp/kdp-niche-scorer-mcp
Pricing tiers
- Free — 5 requests/day
- Pro $9/mo — 30 requests/day
- Author $29/mo — 150 requests/day
Known v0.1 limitation
The RapidAPI `/search` endpoint doesn't return BSR per-product, so per-competitor revenue estimates are `null` and the score uses the review-count-based demand fallback (built into the formula). v0.2 will add `/product-details` enrichment for per-ASIN BSR.