Skip to main content

Getting Started

Yagr starts with a simple loop:

  1. Install @yagr/agent
  2. Run yagr onboard
  3. Connect the agent to its current orchestrator, model, and optional messaging integrations
  4. Run yagr start
  5. Choose the Web UI or the TUI, then optionally add Telegram later

Under the hood, that setup is not the product goal. It is the bootstrap that lets Yagr do its actual job: translate intent into workflows and operate them over time.

Standard install​

For normal product usage, install Yagr globally:

npm install -g @yagr/agent@latest
# or: pnpm add -g @yagr/agent@latest

Then run the onboarding flow once:

yagr onboard

After onboarding, start the runtime with:

yagr start

What onboarding covers​

yagr onboard is the standard first-run entry point. Internally it drives the same setup flow as yagr setup and configures three things:

  1. Your current orchestrator connection: today that means an n8n instance, API key, project, and local sync folder
  2. Your default LLM: provider, model, API key, optional base URL
  3. Your optional messaging integrations: for example Telegram

Yagr persists this state in the Yagr home, so you do not need to recreate setup per repo.

Runtime home​

User-facing Yagr defaults to ~/.yagr.

This is where Yagr stores:

  • local runtime config
  • n8n bootstrap state
  • generated workspace assets
  • linked gateway metadata

This matters because Yagr should remember the operational context around the workflows it creates. The runtime home is not the memory product; the workflows themselves are. But the home is what makes the agent stable between sessions.

After onboarding​

The main commands you will keep using are:

yagr start
yagr start webui
yagr start tui
yagr gateway status
yagr telegram onboarding

Continue with:

If you are contributing from the repository instead of using the published package, use the development flow documented in the root README.