Quick Start
Set up IAS and run your first AI agent in about 15 minutes. This guide walks you through account creation, repository connection, and agent execution.
Quick Start
This guide takes you from zero to a running AI agent in about 15 minutes. By the end, you will have a workspace in the IAS Command Center, a connected repository, and a local agent executing work against your codebase.
Prerequisites
Before you begin, make sure you have:
- A GitHub account -- used to sign in to the Command Center and for Git operations.
- Node.js 18+ -- check with
node --version. IAS requires Node.js 18 or later. - A Git repository -- any repo you want to connect. It can be an existing project or a fresh repo.
- Git configured -- with SSH keys or HTTPS credentials for pushing to your remote.
Step 1: Sign up
Go to app.ias.dev and sign in with your GitHub account. This creates your IAS account and takes you to the Command Center.
If your organization already has an IAS account, ask your admin to invite you to the existing workspace instead of creating a new one.
Step 2: Create a workspace
Workspaces are the top-level container in IAS. They hold your repositories, goals, agents, and team members.
- In the Command Center, go to Settings in the sidebar.
- Click Create Workspace.
- Enter a short, memorable name (e.g.,
my-teamor your organization name). - Click Create.
You will select this workspace when you authenticate the CLI in a later step.
Step 3: Add a repository
Connect a Git repository to your workspace so IAS knows what codebase to work with.
- In the Command Center, go to Repositories in the sidebar.
- Click Add Repository.
- Enter your repository's clone URL:
- SSH:
git@github.com:your-org/your-repo.git - HTTPS:
https://github.com/your-org/your-repo.git
- SSH:
- Click Create.
The repository now appears in your workspace. The Command Center will show it as awaiting agent connection until you start a local worker.
Step 4: Install the IAS CLI
Install the IAS command-line tool globally:
npx @resonancelabsai/ias setupThe setup command does three things:
- Creates a configuration file at
~/.ias/worker.json. - Walks you through connecting to your Command Center deployment.
- Optionally runs authentication (next step) in the same flow.
If you prefer to separate installation and authentication, pass --no-login:
npx @resonancelabsai/ias setup --no-loginAfter setup, you can run all subsequent commands with npx @resonancelabsai/ias <command>, or install globally:
npm install -g @resonancelabsai/iasWith a global install, all commands below simplify to ias <command>.
Step 5: Authenticate
If you skipped login during setup, authenticate now:
npx @resonancelabsai/ias auth loginThis opens the Command Center in your browser. Sign in, select your workspace, and approve the CLI. The token is stored locally at ~/.ias/auth.json.
Verify your authentication status at any time:
npx @resonancelabsai/ias auth statusStep 6: Link your repository
Tell the CLI which local checkout corresponds to the repository you added in the Console:
npx @resonancelabsai/ias repo link /path/to/your/repoThis command:
- Registers the local path in your
~/.ias/worker.jsonconfig. - Upserts the repository in the control plane so the Console can match it.
- Optionally offers to bootstrap IAS into the repo if it is not already installed.
You can link multiple repositories. Each one gets its own entry in the config.
Step 7: Bootstrap IAS into your repository
Bootstrapping adds the IAS scaffold to your repository -- documentation structure, agent configs, and context templates. This is what enables agents to work intelligently with your codebase.
npx @resonancelabsai/ias install /path/to/your/repoThe interactive installer walks you through:
- Profile selection -- choose
control-plane(default, Console-managed),hybrid(adds automated runner), orfull(everything). - Git operations -- creates a branch (
ias/bootstrap), commits the scaffold, and optionally pushes and opens a PR.
After bootstrap, your repository will have a docs/ias/ directory with project context, decision records, and agent process documentation. See Deploying IAS to a Repository for details on profiles and options.
Step 8: Start the agent
Start the local agent so it can receive and execute tasks from the Command Center:
npx @resonancelabsai/ias agent startThis starts two components:
- The setup agent (worker) -- handles coordination tasks like bootstrapping, context building, and applying decisions.
- The work agent (runner) -- handles AI coding tasks and PR reviews.
You should see output like:
[ias] Starting setup agent (worker) + work agent (runner)...
[worker] Starting control-plane polling...
[worker] Heartbeat sent. Status: online
[worker] Polling for jobs...Keep this terminal open while you work. The Console will show your agent as Online.
To run the agent in the background (macOS and Windows):
npx @resonancelabsai/ias worker start --daemonStep 9: Create a goal
With your agent running, create your first piece of work:
- In the Command Center, navigate to your repository.
- Click Create Goal.
- Describe what you want in plain language. For example: "Add input validation to the user registration form" or "Refactor the database module to use connection pooling."
- IAS will refine your intent into a structured goal with scope, constraints, and a readiness scorecard.
- Review and approve the goal.
Once approved, the goal decomposes into tasks. Your local agent claims tasks from the queue and starts executing.
Step 10: Watch the agent work
Monitor the agent's progress in the Command Center:
- Workboard -- see active tasks and their status (pending, running, done).
- Inbox -- answer any decision requests the agent creates when it needs your guidance.
- Repository view -- see commits and pull requests the agent produces.
The agent works in Git branches, produces real commits, and can open pull requests. Review the output as you would any other code change.
Verify your setup
Run the diagnostic command to confirm everything is configured correctly:
npx @resonancelabsai/ias doctorThis checks:
- Node.js version
- Git and GitHub CLI availability
- Configuration file presence and validity
- Control plane connectivity and authentication
- Repository mappings
If anything is misconfigured, doctor will tell you exactly what to fix.
What's next
You are set up and running. Here are the natural next steps depending on what you want to do:
| Goal | Guide |
|---|---|
| Understand the core abstractions | Core Concepts |
| Learn about execution modes (interactive, managed, automated) | Execution Modes |
| Configure agent behavior and policies | Set Up Agents |
| Review and respond to agent decisions | Inbox |
| Add more repositories | Add a Repository |
| Understand the architecture in depth | Architecture |
| Add research tools (MCP servers) | Research Tools |
| Troubleshoot issues | Troubleshooting |
What is IAS?
The Intuitive Agent System (IAS) is a control plane for AI-assisted software development. Orchestrate coding agents across your repositories with full auditability, human-in-the-loop decisions, and local-first execution.
Core Concepts
Understand the key building blocks of IAS -- workspaces, repositories, agents, jobs, goals, the Context Lake, and the principles that tie them together.