Ato turns GitHub repositories and local projects into runnable recipes — with runtimes, dependencies, permissions, and launch settings captured together.
Works on the maintainer's machine. Fails on yours. Setup instructions are stale before they're read.
Node 20, Python 3.11, pnpm, uv, Postgres — all implicit. One missing version and it breaks.
.env, API keys, ports, and network access are discovered too late — usually when something doesn't start.
Even if the code builds, the actual command, env, and permissions that worked are never recorded.
Give Ato a GitHub repo, local folder, or shared recipe. Ato reads what the project needs, prepares the missing runtime pieces, and launches the app on your machine.
Use a GitHub repo, local folder, or shared recipe link.
It finds the runtime, commands, ports, services, and required environment.
It resolves the missing tools and dependencies without global setup.
Ato starts a controlled local session and records the launch for replay.
The next run is not a guess. Ato keeps the launch recipe so the same project can be run, shared, and inspected again.
A Capsule is Ato's shareable launch recipe. It keeps source readable while capturing the runtime, dependencies, entrypoint, policy, and identity needed to launch locally.
Start with real apps. Each recipe captures the runtime, dependencies, launch command, ports, and permissions needed to run it locally.
Web UI
Skip local Node setup.
Notes
Run a self-hosted notes app locally.
Workspace
Try a full workspace app without setup.
Local AI
Connect local AI tools safely.
Recipe gallery coming soon
Docker packages images. Nix pins builds. Dev Containers define workspaces.
Ato captures the launch.
Packages images.
Best when the app is already containerized.
Pins builds.
Best for reproducible packages and systems.
Define workspaces.
Best for editing code in a prepared environment.
Captures launches.
Best for running source apps locally without setup.
Ato works with the tools and stacks you already use. Skip the deployment tax, reduce cloud costs, and keep your workflow flexible — whether you're shipping a prototype, an internal tool, or something in between.
No idle preview servers or temporary managed services.
Skip domains, SSL, deploy hooks, monitoring, and cleanup.
Don't expose half-finished prototypes to the public internet.
Use CLIs, desktop UIs, scripts, databases, or local AI tools.
Ato is pre-1.0. Use it for trusted repositories and local experiments while sandboxing and network controls continue to mature.
GitHub →
Install Ato, run a repository, then inspect the generated recipe and lock data when you want to make it repeatable.
