About AIdeaPulse
AI-powered startup idea discovery, updated daily.
How it works
Every day, AIdeaPulse scans 8 sources for demand signals — Reddit, Hacker News, Product Hunt, GitHub Trending, Dev.to, Lobste.rs, NewsAPI, and Google Trends. A two-stage AI pipeline classifies each signal, then produces a structured idea brief with market sizing, competitors, build complexity, and a confidence score.
The pipeline
- Scrape 600+ signals daily from 8 sources
- Pre-filter by engagement (top ~65 signals)
- Stage 1: Classify with AI — pass or skip
- Stage 2: Full analysis — idea brief with market data
- Deduplicate and store in the database
What you get
- Structured idea briefs with TAM/SAM/SOM market sizing
- Competitor analysis with specific named competitors
- Build complexity and timeline estimates
- Monetization angle with pricing suggestions
- Confidence score (0-100) based on signal strength
- Source links back to the original discussion
Built with
Python pipeline on Linux, Claude API for analysis, Cloudflare Workers + D1 for the API, Astro + React for the frontend. All deployed on Cloudflare's edge network.