API & export
The dataset is the point of this project, so it is open. Read-only JSON, CSV on the same endpoints, no key required.
Endpoints
GET /api/v1
Lists every endpoint.
GET /api/v1/stats
Platform totals and the headline accuracy numbers.
GET /api/v1/topics?status=open|closed|resolved&limit=50&offset=0
Public topics with their resolution source and outcome. Drafts are never included.
?format=csv
GET /api/v1/topics/{slug}
One topic with its options. The crowd split is withheld until the topic closes.
GET /api/v1/calibration
Crowd confidence against what actually happened, in deciles.
?format=csv
GET /api/v1/dilemmas
Every published dilemma with its national split.
?format=csv
GET /api/v1/price-games
Revealed price games: real price, crowd median, error.
?format=csv
GET /api/v1/leaderboard?board=record|wins
The current season's board, with the same privacy rules the site uses.
?format=csv
CSV downloads
Notes
Aggregates only. No endpoint exposes how an individual voted, and the crowd split on an open topic is withheld so published data cannot anchor the people still voting.
Add ?locale=en|ro|ru for translated titles and labels. Missing translations fall back to English.
120 requests per minute per IP. Responses are cached for a minute.
Free to use with attribution to banutz.md (CC BY 4.0).