Launching today
WikiFix for Confluence
Find and fix issues in your knowledge base
37 followers
Find and fix issues in your knowledge base
37 followers
WikiFix monitors your KB for issues like contradicting pages, duplicates and orphans, and fixes them in one click once you approve, so your team can trust what it reads or hears from AI.
Interactive
Free Options
Launch Team / Built With
Hey Product Hunt! Rezgar from the video here.👋
After heaps of experimenting and benchmarking, I'm excited to bring you WikiFix! Big thanks to @fmerian and @oobydrnk for getting this launch off the ground!
If you've ever handled an incident with a broken runbook, failed audit due to stale compliance docs, or onboarded someone with outdated guides, you may appreciate a way to keep those docs clean and consistent without frequent manual reviews.
WikiFix monitors your spaces for issues - in v1 it's contradicting claims, duplicate and orphan docs. It asks you which answer is right, and writes a fix. You approve every edit and undo it in one click.
That's v1, which only looks inside Confluence. Next could be checking docs against code, configs, or Notion - we'll work out what matters most together.
Before launch I scanned a few publilc spaces (https://wikifix.ai/examples/). Even carefully maintained public docs from brilliant teams are not immune - NASA (57 conflictcts across 732 docs), GitLab (332 in 4,576), Sourcegraph (86 in 901), Kafka (168 in 1629)... E.g. Europa Clipper's electronics vault is made of: aluminum-zinc alloy or titanium/aluminum? 🛰️
Got a space you'd bet is fully consistent? Make it your first scan. If WikiFix finds nothing useful there, tell me, and get your credits back twice over.
I'm around all day. Ask me anything!
FAQ:
It's AI, so some findings are wrong: about 1 in 5 when we hand-checked them, depending on the wiki. WikiFix has a feedback button, and that's what we tune on.
Rovo (or any AI assistant for that matter) can help with targeted pages and problems, but not in spotting issues across the board. They start from your question and fetch the few pages likely to answer it. While finding contradictions starts from no question and puts every claim next to every other claim on the same topic or relevant code.
WikiFix starts out as a Confluence app but will go standalone in v2. Tell us if you want your KB platform supported.
@rezgar Asking which answer is right and then writing the fix is the part I like. Most tools just flag a problem and leave you with a to-do list. And one-click undo makes approving edits feel safe.
@mishaal_rashid Agreed. Today, when AI is good at drafting edits, there's no reason we can't make admin's life a little easier, making the routine review a 5 min click-through activity instead of 1 hour of manual editing.
You also have 2 layers of safety here - WikiFix's 1-click revert and Confluence's own page versioning.
Can the scan run on a schedule, like weekly, or is it manual every time? Also, checking docs against code would be my vote for what comes next
@saverio_donati Yep, scans can run ad-hoc or on schedule. We want you to get all the value without spending time in the app.
You choose the spaces to keep an eye on and forget it, until the weekly/monthly report comes in with the new findings. You click through the report, answer some questions or send them to owner, and forget about it until the next report.
Your vote is noted, thank you! That's the part we're also excited about - in my career, I've seen countless runbooks, onboarding procedures, set up guides - all extremely useful, until they stop working midway. Is that the basis for your suggestion too? Or did you have other scenarios in mind?
How does WikiFix decide which page is actually the correct one?
WikiFix doesn't judge. It presents you contradictions as multiple-choice questions like "Who may sign off an exception for a High or Medium risk?". With options and quotes.
If you know the answer, you choose it and WikiFix updates the relevant pages. If not - you can send it to Page/Space owner to decide.
Regular users browsing Confluence will also see a badge saying "X issues found" and can apply an answer if they know it.
Appreciate that you put the "1 in 5 findings are wrong" part right in the post. Most launches skip that kind of thing, so it made me trust the rest more
@camille_messer Thank you, Camille. Unfortunately, even with the best models, you can't weed bad apples out completely.
That said, a big part of WikiFix is the quality gate, vetting and scoring the findings before bringing them to your attention. Nobody wants to dig through AI slop for a rare gem. We only show you the best ones by default, but you can also expand to less reliable ones if you want, which might be useful in critical contexts like legal, medical, compliance...