Ship a support agent
Ground a chat assistant in your own docs with retrieval on every turn.
Attach an inference provider once and route every request — chat, image, speech — through a single project key.
Start from a working example and swap in your own provider keys.
Ground a chat assistant in your own docs with retrieval on every turn.
Batch catalogue art from a prompt file with a diffusion endpoint.
Turn release notes and articles into audio with a speech model.
Adding a provider enables every service that provider supports.
Language models with live web lookup wired into the request path.
An inference host that specialises in diffusion and video models.
Serverless GPU runtime built for AI and data teams.
A cloud for running open-weight generative models at scale.