Skip to main content

​
Action Llama

Run agents like scripts: triggered by cron or webhooks.

Quick Start

Create your first agent in minutes

CLI Reference

All available CLI commands

Configuration

Project and agent configuration

Examples

Ready-to-use agent configurations

​
How it works

  1. An agent is triggered by a webhook, cron schedule, or another agent
  2. The agent runs according to the instructions you define in ACTIONS.md
  3. The agent shuts down

​
Key features

  • Contained runs — each agent run is isolated in a Docker container with only the credentials it needs
  • Git-native — define agents in a git repo, add custom ones, and share them
  • BYOM — bring your own model (Anthropic, OpenAI, Groq, Google Gemini, xAI, Mistral, OpenRouter, or any custom provider)
  • Deploy anywhere — run locally, on a VPS, or in the cloud (AWS ECS, Google Cloud Run)

​
Get started

​
1. Create a project

npx @action-llama/action-llama@next new my-project
cd my-project

​
2. Create and manage agents

npx al chat
The chat console helps you create and manage agents. You can also create agents manually — see the creating agents guide.

​
3. Run

npx al start
If any credentials are missing, it will prompt you for them. Credentials are stored in ~/.action-llama/credentials/ — see credentials for details.