Service Reliability

You can subscribe to updates on our system's current health here.

Service Guarantees & Reliability

The OctoAI compute service is currently in Public Beta. This means we are quickly adding features and improving the service based on feedback and learnings from our early customers. Below are what you can expect from us during this period:


The OctoAI service and the APIs (e.g. invoking inferences on your endpoints in the domain) are expected to be relatively stable and are unlikely to see compatibility breaks during Public Beta. These will only be done with careful consideration and communicated to users.

The OctoAI tooling (such as the octoai CLI and libraries such as the Python octoai library) are expected to see more rapid iteration during Public Beta, and will likely have compatibility breaks. As such, they are versioned as 0.x releases, and we will maintain compatibility once we reach 1.0.


We take serving traffic to your endpoints very seriously—these are intended to power your applications after all—and we expect production-worthy uptimes.

However, we are not making any specific reliability guarantees while OctoAI is in Public Beta. If you have a need for particular guarantees, we encourage you to reach out to our Customer Success team so that we can work with you on them. We are available any time at [email protected], on Discord, or through the chat bubble on