Introducing Custom Servers: Deploy production-ready model servers from Docker images

Our new Custom Servers feature lets you deploy production-ready model servers directly from Docker images using just a YAML file.

We built Truss to simplify the process of deploying AI models in production. But sometimes deploying Docker images is the most direct path—such as when using custom in-house images or open-source servers like vLLM.

Custom Servers offer a codeless way to deploy your Dockerized model servers while leveraging all the infrastructure optimizations Baseten provides.

✕

Check out our launch blog for more information, or dive into the docs to get started!