Skip to main content
Open Notes helps your community moderate content in two ways:
  1. AI moderation catches clearly harmful content right away.
  2. Community review lets members weigh in on moderation decisions — especially when the AI isn’t sure, or when you think it got something wrong.
No moderation decision is permanent. Every automated action can be reviewed and reversed by the community.

​
What you’ll see on posts

When the AI flags a post but isn’t confident enough to act automatically, the post stays visible with an “Under Review” banner. The community decides what to do from there.

​
Participating in community reviews

​
Who can vote

Your forum administrator sets the minimum trust level required to participate. Typically you need to be at least a Member (trust level 2). You earn trust levels through normal participation — reading posts, receiving likes, and contributing to discussions. If you can’t see the voting interface on an item, ask your forum admin what trust level is required.

​
The review page

Visit /community-reviews to see items waiting for community input. Each item shows:
  • The flagged post and its surrounding context.
  • The category the post belongs to.
  • The reason it was flagged.

​
How to vote

For each review item you can cast one vote:
  • Helpful — the moderation action (or proposed action) is appropriate.
  • Not Helpful — the content is fine and shouldn’t be acted on.
You can vote only once per item, and votes cannot be changed after submission.

​
How consensus works

Open Notes uses a bridging-based algorithm adapted from Twitter/X Community Notes:
  • Votes from people who usually disagree with each other carry more weight than votes from people who always agree.
  • This makes it harder for any single group to dominate moderation decisions.
  • Consensus reflects genuine cross-perspective agreement, not majority rule.
Once enough votes are in and a clear consensus emerges, the agreed-upon action is applied (or the post is cleared).
While voting is in progress, vote counts and scores are hidden. This is intentional — it prevents bandwagon effects and encourages independent judgment. Results become visible after consensus is reached.

​
Frequently asked questions

Yes. That’s why every automated decision goes through community review. If the community decides the AI was wrong, the action is reversed and the post is restored.
Depending on your forum’s configuration, flags may be sent to Open Notes for community review in addition to the normal staff flag queue. Your flag could be reviewed by the community, not just staff.
Posts that are restored after an overturn are marked as scan-exempt. Minor edits won’t trigger re-classification. If the post is substantially rewritten, it may be classified again.
Staff can see individual votes and tallies. Other community members see only the final consensus result, not individual votes.
Items stay in the review queue until enough votes are cast. Staff can also take action directly on any item without waiting for community consensus.