AI-native deployment workflow

Your AI already writes code. Automaton ships it.

Install the Automaton CLI, connect your AI coding workflow, and deploy real applications to your self-hosted Automaton server. Install the server first, or use the API URL supplied by your administrator.

Install the server first

The standalone CLI needs no .NET; only the NuGet tool requires the .NET 10 SDK or later. The server installer does not require a separate .NET installation.

Canonical workflow

$ curl -fsSL https://github.com/rzvdaniel/automaton/releases/latest/download/install.sh | sh

$ automaton login --server https://api.your-domain.com

$ automaton new "My App"

✓Authenticated with Automaton Studio

✓Generated AI guidance for coding agents

✓Prepared deployment workflow for publish and runners

# Then ask your AI: Build a live ISS tracker with Earth visualization and realtime orbital data.

$ automaton deploy

Install once. Then just describe the app.

Setup teaches your AI how Automaton apps work. After that, you describe the product — the agent handles structure, deployment, and publishing.

Install the CLI

Download one standalone file, no .NET needed. On Windows use the PowerShell script, or the .NET tool if you have the SDK.

curl -fsSL https://github.com/rzvdaniel/automaton/releases/latest/download/install.sh | sh

Connect to your instance

Connect your workspace and give your AI agent local guidance for building Automaton apps.

automaton new "My App"

Ask for the app

Now the prompt becomes the product idea — not deployment instructions.

"Build a live ISS tracker with Earth visualization and realtime orbital data."

The workflow layer between AI-generated code and deployed software.

Automaton CLI connects local AI tools, Automaton Studio, publishing, and optional server execution.

AI AgentGenerates and edits the app locally.

Automaton CLIApplies structure, metadata, and deployment commands.

Automaton StudioManages apps, deployments, publishing, and visibility.

Published AppGoes live as public or private.

RunnerExecutes configured server logic locally or in Docker.

Quickstart workflow.

From local AI workflow to published app.

curl -fsSL https://github.com/rzvdaniel/automaton/releases/latest/download/install.sh | sh
automaton login --server https://api.your-domain.com

mkdir iss-tracker
cd iss-tracker
automaton new "My App"

# Then ask your AI:
# "Build a live ISS tracker with realtime orbital data."

automaton deploy

automaton publish

Project structure

Automaton apps stay simple. Editable source and optional server execution are described by the app contract.

iss-tracker/
├─ AGENTS.md
├─ .automaton/
│  └─ app.json
├─ app-source/
└─ automaton.json

AGENTS.md gives coding agents local guidance. Server-side execution is optional and follows the app contract.

Built for modern AI coding workflows.

Automaton works with your preferred coding tools instead of replacing them.

Claude Code

Prompt-driven local coding workflow with Automaton deployment support.

Codex CLI

Generate and iterate locally while Automaton handles structure and publishing.

Cursor

Use your preferred editor and AI coding environment without vendor lock-in.

Gemini CLI

Connect prompt-driven coding workflows to deployable Automaton apps.

Every app needs a Runner.

All apps use a Machine Management Runner. Static sites share its static server; APIs and Actions run in bounded containers.

Install a Runner from your Studio on the server or a separate supported machine.

Learn about runners

Static appsShared static server

Server runtimesRequires Runner

Self-hosted RunnerAvailable today

HostingYour infrastructure

Build and publish your first app in minutes.

Install the server, connect the CLI to your API URL, then tell your AI what to build.