Skip to main content

Autoheal Documentation

A self-improving software factory for the post-coding SDLC. Reliability, security, cost and supportability work that comes after the commit. Build, run, govern and continuously improve agents that run 24x7 in a governed environment, not on individual laptops.

Agent infrastructure​

What every worker agent runs on.

Interfaces​

Work with agents from the terminal or chat.

Integrations​

Connect Datadog, GitHub, Sentry, and the rest of your stack.

View all integrations →

FAQs​

Where does our code and telemetry actually run?

With SaaS, Autoheal runs in an isolated tenant in Autoheal's cloud. With BYOC, the platform and its data run in your own environment, connected or airgapped. See Sovereignty for deployment and model-inference boundaries.

What happens when an agent gets something wrong?

Findings include their supporting evidence so your team can review them. Governance policies can require approval before sensitive tool calls. The self-improvement loop evaluates outcomes and proposes changes for an engineer to approve.

Can we bring our own model?

Agents use models from your organization's approved list. Under BYOC, inference can use your own provider account or model endpoint. See Build Your Own and Sovereignty.

How is this different from a coding assistant?

Autoheal handles the reliability, security, cost and supportability work after a commit. Agents run in a shared, governed environment and read shared context. Autoheal works alongside the coding assistants your developers already use. See How It Works.

How is it priced?

Contact the Autoheal team for current pricing and deployment options for your team.