From discovery to unified access
We plan a unified inference interface for open models. Supported models, specifications, pricing, quotas and service status will be published later. No inference endpoints, API keys or paid subscriptions are currently available.
Available now: catalog snapshot
Read the same public catalog version used by the site at the following endpoint. It returns model metadata and does not perform inference.
GET https://abliter.cn/data/models.jsonThe response contains schemaVersion, updatedAt, modelCount and models. Model fields include slug, name, publisher, family, parameters, formats, downloads and source. Handle null or missing values and cache for at least five minutes.
Updates and subscriptions
Use the RSS feed to follow recent models. The catalog checks upstream every six hours; updatedAt records the actual snapshot time.
Future service architecture
The catalog is maintained separately from future inference services. api.abliter.cn could later provide authentication, quotas and model routing. This page will clearly state availability before launch.