Library / SDK
Inference & Serving
BentoML
1Package 2and 3serve 4models 5as 6APIs
the six Ws · specification
W1
Who
Maintained by BentoML (bentoml).
W2
What
A framework for packaging models and AI apps into standardized services and serving them as APIs.
W3
Where
Runs self-hosted as a Python framework, deployable to containers, Kubernetes, or cloud.
W4
When
Reach for it when productionizing and serving models, including LLMs, with scaling.
W5
Why
Standardizes build, packaging, and deployment with adaptive batching and flexibility.
W6
With
Python and model artifacts; GPUs for compute-heavy models.
W7
Watch
Built by BentoML; open-source (Apache-2.0) with a paid managed cloud (open-core); self-hostable.