Your ideas are ready. Let’s get you set up.

Before you install

You’ll need an agent that can run commands on your computer and a readable source file or clear idea. Model usage belongs to your agent account.

01 Install LemmaReel

Use the lemmareel CLI and lemmareel-install.py installer. Your one-time purchase ($49, regular price $199) activates one computer, with repeat downloads and reinstalls on that computer.

Sign in and open your account. After purchase, create your license and save it somewhere private. It is shown only once.

Download the installer, inspect it if you’d like, and run it with Python. The installer script is public; the product package requires a paid license and device activation. Setup requests your license privately, activates this computer, and verifies the package hash before installing. You’ll also need uv and Git; the installer explains missing prerequisites. The default installs the lightweight CLI. Add --with-render to install the larger Python rendering libraries after system prerequisites are ready.

python3 lemmareel-install.py --server https://lemmareel.com

Add the installation’s bin directory to PATH using the command printed by the installer. Then check the environment:

lemmareel doctor

Follow the missing-dependency instructions. Once the rendering libraries are installed, download and verify the pinned English voice model, then check again:

lemmareel doctor --fetch-model
lemmareel doctor

First-time setup and model downloads require internet access.

02 Connect your agent

Choose your agent and the project where you’ll create videos. This adds a small, managed instruction block. It preserves your existing project rules and backs up changes.

lemmareel init --agent codex --project ./my-video

Open a new agent session in that project. No separate skill or MCP is required.

03 Make your first video

Your agent reads the built-in guidance, creates the storyboard, and runs preview, automatic checks, and the final render. It delivers output/<project>.mp4. No manual approval is required to export.

Provide a local path your agent can read. If a PDF is scanned, supply OCR text or describe the content; LemmaReel doesn’t invent missing pages.

Compatibility, honestly.

From version 0.1.0a3, internet access is required for activation and before each preview, render, and export. Rendering and narration run locally; your chosen AI agent may send source content to its model provider. To move your license to a different computer, contact support first. Your agent also needs its model connection.