What breaks in your knowledge base?

Hey all 👋 Thanks for your support during our first launch on Sunday.

What WikiFix is after is simple: when you open a doc, or ask AI, you should be able to trust what you get. Too often you can't - pages drift, authors contradict each other, and nobody has time to revise pages until it bites.

v1 covers a slice of that: contradicting statements, duplicate and orphan pages, and it's Confluence only.

This of course only scratches the surface of what real-world KBs face. We know about some of the gaps, and some of you pointed at them on launch day:

1. More platforms - WikiFix v2 goes standalone, so it can scan across Notion, Git and other common KB storages - supporting modern and legacy KBs.

2. More sources of truth - runbooks, onboarding guides can agree with every other page and still be wrong about what they describe: the code, the configs, the engineering READMEs, the UI in a screenshot or a guide. @saverio_donati asked for code, @petrkovacik for screenshots.

What we can't tell from here is which of these costs the most in practice, or whether the bigger problems are somewhere we haven't looked yet. If you've lived through it in your own knowledge base, let's discuss!