I run SEOAI, a small SaaS, and I kept hitting the same wall with its customers.
Feature requests came in by email and chat. I tracked them on a board, wrote a changelog when I remembered, and the docs were always one release behind. Customers asked about things that had shipped weeks earlier. Every tool I looked at covered one piece: a feedback board here, a changelog there, docs somewhere else. None of them connected shipping a feature with updating the docs.
So I built duxloop.
Users post ideas on a public board or through a one-line widget in your app, vote and follow.
AI sorts new posts into categories and flags duplicates.
Your statuses are the roadmap. Move a post to Planned or In progress and the public roadmap follows.
When you ship, you write a changelog entry and link the posts. Everyone who followed them gets an email.
Then duxloop drafts the matching docs update. You review it and accept it or dismiss it.
The docs site has semantic search, an Ask AI box that cites the pages it answers from, and llms.txt for AI agents.
SEOAI's customer portal runs on it (support.seoai.de), so I use it every day.
Free plan for small projects. Pro is €29/month for the whole team: unlimited seats, custom domain, white-label, and English/German auto-translation. No per-user or tracked-user fees. Hosted in Germany. Every workspace starts with a 14-day Pro trial, no credit card needed.
Does "the loop ends in your docs" make sense as the main idea, or does it sound like one feature too many?
What would stop you from moving off Canny, Featurebase or a Notion page?
You can try it at https://duxloop.com. Happy to answer anything below.
Elias