How Runo environments work

Runo connects a Git branch to a remote environment on AWS. Your repository recipe defines how the application starts and which services it needs. The interactive diagram illustrates the flow.

Create the environment

Run runo new checkout-fix from a repository with a reviewed and committed recipe. Runo creates the task branch and worktree, then provisions its remote environment. Database services and seed data depend on the recipe.

Work with your agent

Enter the worktree printed by Runo, then run runo agent claude or runo agent codex. The agent runs on the remote VM alongside your application's services. You provide the agent credentials. See coding agents.

Open the application

Run runo url --open from the environment's worktree. Inspect the application yourself or share the URL. The preview does not add authentication; use development data and credentials.

Validate and bring changes back

Run runo validate to execute the recipe's checks remotely. Results and logs are downloaded to the local worktree. Use runo pull to download remote code changes, then save them before destroying an environment.

When you intend to publish a branch, runo ship can commit, push and open a PR. Configured checks and human review still determine whether the change is ready.

The diagram is an explanation, not a live view of infrastructure. The quickstart covers setup and cleanup. AWS resources and agent usage can incur charges.