Developer documentation
Everything you need to connect a repository, choose where the analysis runs and integrate Loopback with your app and your tools.
- QuickstartFrom signup to the first diagnosis on your own code in less than 10 minutes, without writing YAML.
- How it worksThe pipeline from ticket to release, the states of a ticket, the autonomy levels, the approval and the rules that no setting changes.
- .loopback.ymlReference of the configuration file: every key, its values and its default.
- Execution modesManaged, your runner or your agent through MCP: where the code runs, who pays the model, and which agents work in each mode.
- GitHub ActionRun jobs in your GitHub Actions: inputs, outputs, OIDC, permissions and the security model.
- Self-hosted runnerThe runner container: the same harness as the Action, polling over outbound HTTPS only.
- MCP setupConnect your IDE or CLI agent to the remote MCP server: a token, the configuration for each client and the tools.
- WidgetInstall the report widget on the web, React Native, iOS and Android, and how it handles consent and masking.
- Status pages, changelog and badgeThe public page of each ticket, the public changelog, the badge and custom domains.
- Logs and webhooksConnect Sentry, Datadog, CloudWatch or your own endpoint: errors feed diagnoses and the proactive mode.
- Security and data handlingWhat leaves your infrastructure in each mode, how data reaches the model, tokens, retention and how to report a vulnerability.
- REST API referenceEvery public endpoint of the API, generated from its OpenAPI document.