Deploy models

Deploy models as REST API endpoints with a single click for asynchronous (batch) or synchronous (real-time) predictions on new data. Domino offers built-in security, scaling, and management features:

• Monitor endpoint performance.
• Endpoint security.
• Endpoint scaling and routing.
• Endpoint health and logs.
• Endpoint authorization mode.
• Asynchronous endpoints.

In addition, you can score large batches of data with batch scoring Jobs. To serve a Large Language Model (LLM) instead, see Host an LLM, which deploys it as an endpoint with an OpenAI-compatible API.

​

Deploy models to third-party platforms

Domino gives you the flexibility to use your models and endpoints on other supported platforms.

• Deploy models to SageMaker to leverage auto-scaling and streaming capabilities.
• Export to NVIDIA Fleetcommand for edge deployments.
• Export models to Snowflake to bring the model to where your data lives.

​

Integrate Domino models with existing CI/CD workflows

Domino can deploy natively, export to third party platforms, or integrate model deployments with existing CI/CD workflows with the Model Export API. The choice is yours.