Skip to content
Podgrove
Two local worktrees sync to two separate Docker engines in an approved Kubernetes namespace. Each engine contains its own Compose services and persistent volume.Two local worktrees sync to two separate Docker engines in an approved Kubernetes namespace. Each engine contains its own Compose services and persistent volume.

Your Compose stack.
Every worktree.

Keep your editor local. Run a separate Docker engine in Kubernetes for each worktree, with your Compose services and data.
Get startedView on GitHub

macOS & Linux Your existing Compose files No local Docker daemon needed

Keep the workflow.
Separate the stacks.

Another branch shouldn’t mean another stack competing for your laptop’s memory, or a shared database that someone else’s tests can change.

How Podgrove works
Compose stays Compose
Keep services, networks, volumes and healthchecks in your Compose files. Add podgrove.yml for the cluster target and runtime settings.
Each worktree gets its own engine
Separate Docker networks and persistent storage, even when worktrees share one Kubernetes namespace. Set CPU, memory and storage per worktree.
Your development loop stays local
Edit source on your machine and use forwarded localhost endpoints. A background session manages source sync and connections.

Start with your Compose files

Use an approved development cluster and an existing namespace. An administrator prepares access, storage, networking and admission for the privileged engine.

  1. Install Podgrove

    Install a verified release into a versioned environment. The guide covers the required CLI tools and upgrades. Homebrew is not available yet.

  2. Prepare your worktree

    Keep the original Compose files. Set the explicit context, namespace and storage class in podgrove.yml, then generate the administrator’s bootstrap manifests.

  3. Start and verify

    After access is ready, validate the project, start the stack, and run your own tests against its local endpoints. Check that a source edit reaches the running service.

From the application worktree

With PODGROVE_BIN set by the installation guide and namespace setup complete:

"$PODGROVE_BIN" validate
"$PODGROVE_BIN" doctor
"$PODGROVE_BIN" up
"$PODGROVE_BIN" web

Find your next step

Practical guides for developers, agents and the platform team.

Configure a worktree
Resources, persistent storage, namespace modes and optional node placement.
Open the dashboard
A read-only view of services, endpoints, storage, configuration and live logs.
Connect services
Declare access to a local TCP service or a service in another worktree.
Prepare namespace access
Review and apply the generated manifests with the platform owner.