- the loudest users dominate inboxes / DMs
- real sentiment is scattered across places like Reddit, forums, comments, reviews, etc.
Feedstack tries to create a “zero point of truth” for prioritization by combining:
Discovery: automatically find relevant conversations about your product across the web
Distillation: use AI to turn messy threads into clean, review-ready feedback items (title, summary, evidence, suggested feature)
Validation: publish the best items to a public voting board so you can see what has real demand (not just what one person asked for)
How it works (quick)
You enter your product name + site
Feedstack uses SerpAPI to find discussions (e.g. Reddit + forums etc.)
OpenAI extracts the actual requests, groups duplicates, and produces a prioritised list
You can approve items and push them to a public board for voting
What’s different vs “classic” feedback tools
Most tools start with “collect feedback from your users”. Feedstack starts earlier: what are people already saying when they’re not in your app, not on your forms, and not trying to be polite? Then it adds a second layer (voting) to validate.
Current state
Working MVP, used on my own projects + a few early testers
Still rough in parts — especially clustering/duplication and source quality filtering
Link
feedstack.app (happy to share a demo board if anyone wants)
What I’d love feedback on
Does “discovery → distil → vote” match how you’d actually prioritise?
Which sources would you want first (Reddit, GitHub issues, app reviews, Twitter/X, etc.)?
Any obvious privacy/ethics footguns I should address up front?
Thanks — I’ll be around in the comments.