SuperPlane
SuperPlane is an open source automation engine for AI-driven engineering.
It lets you orchestrate engineering workflows across the tools you use — such as Git, LLMs, CI/CD, observability, incident tools, and infrastructure — with durable execution, approvals, and operational UI.
SuperPlane executes your processes deterministically, providing the exact guardrails both humans and AI need to safely interact with your systems.

Project status

SuperPlane is in beta. Self-host the core engine (installation guide) or use SuperPlane Cloud for managed runners and one-click app installs. Core primitives and integrations are maturing; breaking changes are possible. Report issues on GitHub.
What it does
SuperPlane orchestrates your existing stack into git-backed apps with durable execution — workflows too complex for a single script or CI job.
- Apps: A deployable unit combining a workflow graph, custom console UI, app-scoped memory, and deterministic execution. Versioned in git (
canvas.yaml, console.yaml); defines guardrails for AI agents and human operators.
- Event-driven orchestration: Multi-step workflows across your Git, CI/CD, observability, incident tools, and notifications — triggered by webhooks, schedules, and tool events, with approvals, policy checks, and human-in-the-loop steps.
- Console dashboards: Define your own per-app operational UI as a dynamic grid of panels. Use it to display KPIs, tables, charts, runbooks, pinned nodes, and workflow controls, backed by live data from memory, runs, and executions.
- Agents & operators: Built-in per-app agent to design workflows and debug runs; CLI and skills for external coding agents. Same RBAC on all paths.
How it works
- Canvases: A graph of steps and their dependencies; a single canvas can express multiple workflows and run them concurrently.
- Components: Each node is a trigger or action, built-in or integration-backed that performs a specific task (for example: deploy a service, open an incident, post a notification, wait for a condition, require approval, etc.).
- Events & triggers: Incoming events match triggers and start runs with the event payload as input.
- Runs & durable execution: Runs, run items, and payloads are tracked across restarts; failed steps can resume without custom retry logic.
- Memory: App-scoped JSON storage that persists across runs.
Example use cases
A few concrete things teams build with SuperPlane:
- PR preview environments: on pull request, provision an ephemeral environment, run tests, and post the live URL back to the PR.
- Policy-gated production deploy: when CI finishes green, hold outside business hours, require on-call + product approval, then trigger the deploy.
- Progressive delivery (10% → 50% → 100%): deploy in waves, wait/verify at each step, and rollback on failure with an approval gate.
- Release train with a multi-repo ship set: wait for tags/builds from a set of services, fan-in once all are ready, then dispatch a coordinated deploy.
- “First 5 minutes” incident triage: on incident created, fetch context in parallel (recent deploys + health signals), generate an evidence pack, and open an issue.
Quick start
Local (demo container):
docker pull ghcr.io/superplanehq/superplane-demo:stable
docker run --rm -p 3000:3000 -v spdata:/app/data -ti ghcr.io/superplanehq/superplane-demo:stable
Open http://localhost:3000.
Cloud: Sign up at app.superplane.com (cloud beta overview).
For a guided first workflow, see the quick start guide.
Supported Integrations
SuperPlane integrates with the tools you already use. Each integration provides triggers (events that start workflows) and components (actions you can run).
View the full list in our documentation. Missing a provider? Open an issue.
AI & LLM
Version Control & CI/CD
Cloud & Infrastructure
Observability