Find and merge duplicate feature requests

Select one request from your feature list and choose Find Duplicates to scan it against other active requests. Scans take about 30–60 seconds, depending on the number of requests. AI considers the action, technologies, constraints, and context being requested, and shows similarity scores with explanations. Scores of 90–100% are labeled Duplicate, 70–89% Very Similar, and 65–69% Related; matches below 65% are not shown. Treat the suggestions as guidance: related requests may still serve different use cases.

Review the suggested groups, scores, and explanations before choosing which requests to merge. Select at least two requests, then choose Next Step and Confirm Merge. You decide what to merge; a similarity score alone does not determine whether requests are truly the same.

If you already know which requests to combine, select multiple requests on the main board, choose the merge action, and choose which request should remain primary. This does not require running AI detection first.

Choose a primary request, which remains visible on the board. You can optionally update its title and description or generate combined content with AI and edit it before confirming. Votes transfer to the primary request; duplicate votes from the same user are deduplicated. Comments move to the primary request. The other requests are archived, not deleted, and a merge audit comment records which requests were merged. Archived duplicates do not redirect to the primary request, so share the primary request link after merging.

Avoid merging requests that target different user groups, describe a subset that might ship separately, have significantly different timing, or propose conflicting implementation approaches. Similarity scores are a prompt for review, not a decision to merge.