How to get your mitts on the data

Everything on this site is built from data contributed by users, and it's available for you to do whatever you want (modulo the license). The full corpus is exported periodically under CC BY-SA 4.0.

In the latest dump

5 reviews

9 models

6 tasks

Download

modelreviews-latest.zip

Sqlite snapshot of every public table. Refreshed on a regular cadence; the latest dump always lives at this address.

Each archive ships both CSV (for spreadsheets and quick loads) and newline-delimited JSON (for pipelines), plus a README documenting the schema and the export timestamp.

What's inside

models
Every model with its provider, slug, and the date it was added.
providers
The organizations shipping the models, with their model counts.
tasks
The task taxonomy reviews rate against, with slugs and names.
reviews
Approved reviews: title, body, author handle, and timestamp.
ratings
Per-task scores, joining each review to the tasks it rated.
tags
Community tags and the model/review rows they're applied to.

Using it

When you use this data, the only strings attached are attribution and share-alike; see the license for the exact terms. If you build something interesting, we'd love to hear about it via the data inbox.