What is Lifecycle?
Understand how Lifecycle creates connected, isolated application Environments for pull requests and API-driven workflows.
Lifecycle creates isolated application Environments on Kubernetes. It selects the Services that a change requires. It builds the changed sources and deploys the result with its dependencies. Lifecycle publishes useful URLs and cleans up Environments that are no longer necessary.
Pull requests are the usual trigger, but Lifecycle can also create Environments without a pull request through an enabled UI or API workflow. Long-lived static Environments can supply fallback Services.
The core flow
- A repository contains a correct Lifecycle configuration.
- A pull request, UI action, or API request asks for an Environment.
- Lifecycle resolves necessary and optional Services and their source revisions.
- Lifecycle builds the changed sources and deploys the Kubernetes workloads.
- Users examine status, logs, URLs, and actions in GitHub, the UI, CLI, API, or MCP.
- When the owning workflow ends or its lease expires, Lifecycle tears down ephemeral resources.
An Environment is the user-facing unit. Some API and CLI interfaces use the
historical term build or the field name uuid. Those names usually refer to
the same Environment or its human-readable ID.
What Lifecycle helps with
- Fast feedback: Test changes before merging.
- Isolation: Prevent unrelated changes from colliding in a shared staging Environment.
- Connected Services: Use changed Services with configured dependencies and fallbacks.
- Shareable results: Supply developers, QA, and stakeholders with stable links for the lifetime of the Environment.
- Controlled cleanup: Tear down pull-request Environments automatically and expire leased API Environments.
- Agent workflows: When enabled, let users and approved AI tools examine the same Lifecycle state through governed interfaces.
View the demo
Select your next step
- Application developer: Onboard a repository and create your first Environment.
- Evaluator or platform operator: Select an installation path. Use the starter infrastructure prerequisites only for the OpenTofu evaluation path.
- API user: Create a personal API key and create a PR-less Environment.
- Agent user or administrator: Start with Agent Sessions and Agent administration.