Skip to content
LlamaIndex Python Documentation
TypeScript

Introduction to LlamaDeploy v2

LlamaDeploy is still in beta. You can try it out locally, or request access by contacting us

Build and ship small, focused agentic apps—fast. Start from LlamaIndex workflows you already prototyped, iterate locally, and deploy to LlamaCloud right from your terminal.

  • Customizable agentic apps for back‑office and utility use‑cases
  • Rapid iteration; great fit for code‑assistant driven development
  • Run headless (API‑only) or with a bundled UI
  • Powered by LlamaIndex workflows you control
  • First‑class with LlamaCloud for document‑centric apps
  • One‑command deploy to LlamaCloud

Getting Started

Section titled “Getting Started”

This guide walks you through installing llamactl and creating your first local app and deployment.

Prerequisites

Section titled “Prerequisites”
  • UV for Python and dependency management (required)
  • Node.js for UI development (optional unless you add a UI). Use your package manager of choice (pnpm, yarn, or npm).

Install

Section titled “Install”

Choose one:

  • Try without installing:
Terminal window
uvx llamactl --help
  • Install globally (recommended):
Terminal window
uv tool install -U llamactl
llamactl --help

Initialize a project

Section titled “Initialize a project”

llamactl includes starter templates for headless services and full‑stack UI apps. Pick a template and customize it.

Terminal window
llamactl init

This creates a Python module that exposes your LlamaIndex workflows, plus an optional UI you can serve as a static frontend.

Develop

Section titled “Develop”

Run the dev server for your application:

Terminal window
llamactl serve

llamactl serve reads configuration from your app’s pyproject.toml (or llama_deploy.yaml) to discover workflow instances in your Python module. See the Deployment Config Reference for details and examples. For CLI flags, see llamactl serve.

To define and export workflows, see Workflows & App Server API.

To persist and query structured outputs from your workflows or UI, see Agent Data Overview, with guides for TypeScript and Python.

.env files

Section titled “.env files”

Most apps need API keys (e.g., OpenAI). Load them via .env and reference the file in your config:

[tool.llamadeploy]
env_files = [".env"]

Then set your secrets:

.env
OPENAI_API_KEY=sk-xxxx

Create a Deployment

Section titled “Create a Deployment”

Push your code to a Git repository. LlamaCloud will clone, build, and serve your app.

From your project directory, create a deployment:

Terminal window
llamactl deployment create

The first time you run this, you’ll be prompted to log into LlamaCloud. See llamactl auth for more info

Fill in the interactive form (most fields are auto‑detected):

  • Name: Human‑readable and URL‑safe; appears in your deployment URL
  • Git repository: Public HTTP or private GitHub (install the LlamaCloud GitHub app for private repos)
  • Git branch: Branch to pull and build from (use llamactl deployment update to roll forward)
  • Secrets: Pre‑filled from your local .env; edit as needed

After creation, the TUI shows status and logs. Later, use llamactl deployment get to view, llamactl deployment edit to change config, or llamactl deployment update to roll a new version.

Next: Read about defining and exposing workflows in Workflows & App Server API.