Use case
Building taste-graph catalogs with compere
Turn “A or B?” clicks into a ranked catalog that grows with your data.
Who it's for: Product teams building preference-driven catalogs
The problem
You want a ranked catalog that reflects real user taste, built from lightweight A-or-B interactions rather than heavyweight ratings. It needs to start simple and scale without a rewrite.
How compere approaches it
- Register catalog items as entities via the REST API.
- Serve UCB-selected pairs in your UI; record each click as a comparison.
- Elo maintains a live ranking that updates as taste data accumulates.
- Start on SQLite; switch to PostgreSQL with DATABASE_URL when volume grows.
Where it fits (and where it doesn't)
compere fits preference-ranking catalogs where the input is pairwise choices. It ranks entities; recommendation and personalization logic live in your application on top of the ranking.
Features used
Ready to try this?
The quickstart gets you from install to a live leaderboard in minutes.