
Cloud API for running open-source and hosted AI models (image, video, audio, language) without managing GPUs. Client libraries for Node.js and Python plus an HTTP API and webhooks.
CPU (Small) hardware, $0.09/hr; many models are priced per output instead.
Accounts with credit but no payment method: 6 requests per minute.
Local file uploads via client; data URIs recommended only under 1 MB.
API prediction inputs, outputs, files and logs are removed after an hour by default.
Nothing new in the last 30 days. For older changes, see the official changelog.