Beta v0.1.x

Route durable work across machines you trust.

Workplane is a control plane for shell tasks, local inference, and batch agent harnesses—on your laptop, home server, and office GPU box—over Tailscale, WireGuard, or a private LAN.

Quick start View source 
npm install -g workplane
workplane-db-migrate && workplane-server

How it works

Nodes pull work outbound—no inbound ports on workers. You get durable state, logs, artifacts, and retries without SSH session juggling.

1. Submit

Use the CLI to create a task with capabilities (ollama, codex, shell, …).

2. Route

The scheduler assigns work to an online node whose capabilities match requires.

3. Execute

Adapters run in an isolated workspace—clone, branch, capture diffs and logs.

Adapters

Extensible execution—agents and inference are adapters, not the core.

Adapter Kind Typical capabilities
shell shell.exec shell
ollama inference.batch ollama
aider agent.run git, aider
codex agent.run git, codex
claude-code agent.run git, claude-code

Interactive AI client sessions (PTY, stdin over the control plane) are planned for v0.2.0.

Get started

Node.js 20+ and Postgres. Auth tokens recommended outside localhost.

  1. npm install -g workplane
  2. Set DATABASE_URL, WORKPLANE_NODE_TOKEN, WORKPLANE_OPERATOR_TOKEN
  3. Run workplane-db-migrate, workplane-server, workplane-node
  4. workplane task submit shell --command "echo hello"

Documentation