Inference service · Managed inference host

Replicate

Replicate's managed API and deployment platform for running public, private, and customer-packaged machine-learning models.

Operator

Replicate

Delivery

On-demand · Batch · Dedicated endpoint

Model sources

Open-weight models · Customer-supplied models

API styles

Service-native

Boundary

Represents Replicate's hosted prediction and deployment service, not model authors, the open-source Cog packaging tool alone, or deployments run outside Replicate.

Open in the directory → Open official service documentation ↗

Editorial ratings appear in the directory beside the profile they belong to and are never compared across collections.