Compartment
Deploy small internal applications on infrastructure you control.
Deploy and observe Python agent workflows on your own infrastructure

CanyonOS is an open-source platform for deploying and operating applications made of multiple AI agents. Developers write agents in plain Python and use its command-line tools to run a workflow locally or deploy it to their own AWS account or Kubernetes cluster. The platform carries calls between agents so individual parts of a workflow can be observed and scaled independently.
It exposes information such as agent costs and queue times, exports telemetry to existing monitoring tools, and supports policy files for spending and model choices. It works with providers that offer an OpenAI-compatible API. CanyonOS is available under AGPL-3.0; its initial public 0.2 release is an actively developed project that has not yet been proven at large production scale.
Tried it? Share a useful thought or ask a question.
Sign in to join the discussion.
Sign in to comment