Real scraped signals — not LLM brainstorming. Reddit pain volume, Google Trends momentum, and Claude AI synthesis combine into a single Winnability Score for any market.
Scans Reddit and Hacker News for posts expressing frustration, need, or active problem-seeking. Weighted by engagement. More pain = bigger latent market.
Pulls 12 months of Google Trends data and computes the slope. Rising interest means you're early. Flat is fine. Declining is a hard no.
Claude reads the raw signals and outputs a structured Winnability Score with per-component breakdown and concrete action items — not generic startup advice.
NicheIQs is built for autonomous use. Works with Claude, LangChain, CrewAI, AutoGPT — anything that can call an HTTP endpoint or load an MCP tool.
// claude_desktop_config.json { "mcpServers": { "nicheiqs": { "command": "python", "args": ["/path/to/mcp_server/server.py"], "env": { "NICHE_ENGINE_API_KEY": "ne_your_key" } } } }
Two endpoints. Submit a niche, poll for the result. Structured JSON that drops straight into any AI pipeline, LangChain agent, or autonomous workflow.
View full API docs ▸# 1. Submit POST /analyze X-API-Key: ne_your_key {"niche": "AI invoice software"} → {"job_id": "abc123"} # 2. Poll GET /results/abc123 → { "score": 74, "verdict": "strong", "breakdown": { "pain_score": 82, "demand_score": 71, "competition_score": 68, "monetization_score": 75 }, "recommendations": [...] }