ModelFoldOne API for all the best AI models

Video, image, chat and speech models behind a single endpoint, a single key and a single billing unit. This is the English documentation site.

Getting Started

All models (192)

API Reference

One endpoint for every model

Images and video go through /v1/images/generations and /v1/videos/generations; chat uses the OpenAI-compatible /v1/chat/completions. Switching models only means changing the model field.

Metered in credits, refunded on failure

Media models are charged per call. Chat models are settled against the token usage the upstream reports. Failed upstream calls are refunded automatically.

A reference page per model

All 192 models get their own page with family, billing unit, upstream cost and a copy-pasteable request example, regenerated from the model catalog.

English documentation is being written

The English site is scaffolded and mirrors the Chinese structure page for page, but most prose here is still a skeleton. The generated model and pricing pages under Models and Pricing are complete, since they are produced from data.

For the full written documentation, use the Chinese documentation in the meantime.

ModelFoldOne API for all the best AI models

One endpoint for every model

Metered in credits, refunded on failure

A reference page per model

English documentation is being written ​

Start here ​

English documentation is being written

Start here