Duplicates, detected and merged automatically.

Every product team that runs a public feedback board hits the same problem within a few months: the same idea gets submitted over and over, in slightly different words. "Add dark mode" becomes "Please add a dark theme" becomes "Night mode would be great" becomes "Can the dashboard be dark?" Each post collects votes independently, the signal fragments across copies, and your top-requested feature suddenly looks like four different mid-tier requests. Merge AI exists to fix that, automatically, before it corrodes your prioritization.

Merge AI uses semantic search rather than keyword matching, which means it catches duplicates that traditional dedupe systems miss. "Dark mode" and "night theme" are matched as the same underlying request even though they share no words in common. When a duplicate is detected, Upvoty surfaces it in a review queue where a moderator can confirm the merge in one click, or you can configure auto-merge above a confidence threshold so the system handles it without human intervention.

What happens when you merge a duplicate

Merging in Upvoty is non-destructive and reversible. When two posts are merged, votes transfer to the canonical post, comments are preserved and reattached, and every user who voted on either copy is automatically subscribed to updates on the merged result. The deprecated post stays accessible via its original URL with a redirect to the canonical version, so any external link or share continues to work.

That last detail matters more than it sounds. Without a redirect strategy, merging breaks links shared in Slack channels, blog posts, and customer emails. Upvoty preserves those links forever, which means you can dedupe aggressively without worrying about breaking the long tail of references to the original posts.

Merge AI plays well with the rest of your stack

Merge AI gets dramatically more accurate when paired with smart tags. Tags narrow the candidate space, so the model is comparing posts within the same feature area rather than across your entire board. The result is fewer false-positive suggestions and a tighter, higher-confidence merge queue.

Combined with moderation, Merge AI can run at the front door: as soon as a new post enters the approval queue, the system checks for likely duplicates and surfaces them to the moderator in the same screen. The moderator can then approve and merge in one action, ensuring the canonical post grows and the board stays clean from the very first submission.

Cleaner data, sharper roadmap, better decisions

The downstream effect of automated deduplication is everywhere. Your feedback boards stay scannable, your public roadmap reflects genuine demand instead of fragmented copies, and the data exports you push into your analytics warehouse are no longer polluted by near-duplicates that skew aggregate counts. When your CFO asks how many customers requested feature X, you can give an answer you actually trust.

Merge AI also reduces moderation load. Without it, moderators spend a significant chunk of their time recognizing duplicates manually, performing the merge, transferring votes, and writing apologetic comments to users whose post was deprecated. With it, the work is collapsed into a one-click confirmation, and increasingly, into an automated decision the system handles entirely on its own. That is real product-ops leverage, and it compounds with every additional piece of feedback that comes through the system.