Self-Deployed Container
The Extract container lets you run the full document extraction pipeline on your own infrastructure. It packages the same ML models and processing logic as the hosted Extract API into a single container image that you deploy and manage.
Why self-deploy?
- Data residency. Documents never leave your environment. The container processes everything locally with no external network calls, keeping all data within your own network boundary.
- Latency control. No round-trip to an external API. You control where the container runs relative to your application and the network path between them.
- Throughput scaling. Scale horizontally by running more instances to match your workload.
Getting access
Access to the container image is provisioned by Kensho. To request access, please email extract@kensho.com.
Getting started
Follow the guide for the container version you're running: