Prepare Your Dataset
Load your list of records (e.g., company names, user accounts) that you suspect contain duplicates. This can be a large JSON file, a CSV, or a database query result.
Efficiently clean up large databases by identifying and merging duplicate or near-duplicate records. Use a decision model to get confidence scores for potential matches, allowing for automated merging or manual review.
Download the skill for this workflow. Version 1.0.1.

Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
6 steps
Load your list of records (e.g., company names, user accounts) that you suspect contain duplicates. This can be a large JSON file, a CSV, or a database query result.
Feed the list of records to Jev. Configure it to compare entries and identify potential duplicates, such as 'Cedar Grove Office Products' vs 'Cedar Grove Office'.
Test on a representative sample first, then batch candidate comparisons and measure cost and accuracy before scaling.
For each potential match identified, Jev should return a confidence score indicating the likelihood that the two records are duplicates. This score is key to automating the next step.
Define a confidence score threshold to determine what to do with the matches. For example, evaluate a 99% confidence threshold against labeled examples before enabling automatic merges; keep a backup and review uncertain matches.
For pairs with a lower confidence score (e.g., between 75% and 99%), flag them for manual review by a human operator. This creates a human-in-the-loop system that balances automation with accuracy.
For the pairs flagged for review, pass them to a cheap generative model (such as one accessed through OpenRouter). Prompt the model to explain why it thinks the pair is a match, providing a qualitative reason to assist the human reviewer.
Turn an idea into a PRD, user stories, and a plan.

Make your household's priorities and progress visible to everyone by creating a custom dashboard. Use Claude Code to write a script that extracts key information from your household knowledge base and displays it on a digital wall display like DAKboard.

Tackle a complex household project, like comparing insurance quotes, by using your AI-powered knowledge base. Use Claude Code to read relevant documents, identify blocking questions, and draft communications to move the project forward.

Turn static brand character images into animated sprites with multiple poses (e.g., walking, cheering, jumping) for use in marketing videos. This workflow leverages Fable 5.2's creative and geometric capabilities to generate convincing motion for abstract characters.
Join 250,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.