Signup is free, just an email, no payment. I know some people dislike logins, but it helps me prevent spam and lets you save your runs. You can also play with a sample deck if you want to see it work before adding your own.
What I would love from HN is feedback. Are the matches useful? Are the explanations clear enough? Do you think it needs more transparency about how the ranking works, or less? Where does it break? Thin decks and biotech are still tricky, and sometimes investor data gets stale. If you notice bad fits, I would be grateful if you point them out so I can fix the signals.
I will be here in the comments to answer technical questions and to hear your thoughts. If signup is a blocker, let me know and I can share a sample run so you can still take a look.
Veefa•1h ago
Under the hood it’s a mix of LLM parsing and embeddings, with some rules layered in. The data is public and updated regularly. The stack is pretty simple: Next.js, FastAPI, Postgres.
It’s not perfect yet. Thin decks and biotech are tough, and some investor profiles get outdated. I’m working on ways to show uncertainty better.
Would love feedback on whether the matches feel useful, where you see bad fits, and how much explanation you’d want on the ranking. I’ll stick around to answer questions.