Auto-collection
A corpus of ~3100 startups: YC, Product Radar, HackerNews — daily.
Every day the system reads ~3100 startups from YC, Product Radar and HackerNews, runs them past a council of AI analysts, and sends you a short brief: who launched, who stalled, where the market is moving. A longread on the web plus a digest in Telegram. Every number is computed by code, not an LLM — so it can't make them up.
"Which competitor just raised? Who shut down? Where is the market actually heading — and did I miss something that mattered?"
You're heads-down on your own product — and you just can't keep up with everyone else. And you can't close this by hand: more sources every week, hundreds of stories, the thing that matters buried in noise. So you hear the important news last — when there's nothing left to do about it.
A corpus of ~3100 startups: YC, Product Radar, HackerNews — daily.
Agents surface changes, trends, and strategic insights.
Numbers are computed by Python, not an AI; every claim shows its source and a confidence level.
An editorial longread + a top-3 digest with takeaways in Telegram.
~3100 startups from YC, Product Radar, HackerNews — with no manual work.
Strategic insights, not raw data.
A longread + a Telegram digest. Numbers are verified by code — no hallucinations.
The market pulse in your pocket, issue history at hand.
You know your competitors' fate and market trends every day.
Numbers are computed by Python — a source under every claim.
Open Telegram — a top-3 digest + takeaways.
The corpus is collected and analyzed automatically.
On the web and in Telegram. Numbers are computed by code.