Documentation

Guides for setting up and configuring SetForth, whether you run agents in our cloud or on your own machines.

• Overview — Orchestrating Claude Code AgentsSetForth handles the orchestration of background agents that write code and open pull requests — in the SetForth cloud or on your own self-hosted runners.
• Project ConfigurationConfigure a project's sandbox: setup script, Docker Compose services, nested Docker, and execution target, so background agents build, test, and run your code the way you do.
• Automated Pull Request ReviewsReview, give feedback on, and merge agent pull requests — with optional pull request automation that merges only when every acceptance criterion passes.
• Self-Hosted RunnerInstall and register a runner to run background AI agents on your own compute, with your toolchain, private network access, and Claude Code under the hood.
• TroubleshootingFix the most common self-hosted runner and project problems — each written as symptom, likely cause, and resolution — to keep background agents running.
• IntegrationsConnect the tools you already use, so agents check their own work against your real production data before a task counts as done.
  • Sentry
  • PostHog
  • Amplitude
  • Mixpanel
  • Datadog
  • Grafana
  • New Relic
  • Railway
  • LaunchDarkly
  • Custom Connector
• Claude Design ConnectorSend a design straight from Claude to SetForth, where it lands as an initiative with the design files attached, ready to plan and build.