OpenMontage

What it is

A free, AGPL-3.0 open-source system that, per its README, turns your AI coding assistant into a video production studio. There is no code orchestrator: the assistant (Claude Code, Cursor, Copilot, Codex or Windsurf) reads pipeline manifests and skill files and calls Python tools for research, script, scene plan, assets, edit and compose, rendering with Remotion, HyperFrames or FFmpeg. The README says proposal, script, scene plan, generated assets and publish each pause for your approval. You can paste a reference video and it returns concepts, a tool path, cost estimates and a sample. It is not one button, and the assistant is not included.

Setup

1. Check the basics. You need Python 3.10+, FFmpeg, Node.js and an AI coding assistant. The README says Node 18+, but HyperFrames needs 22 or newer, so use 22+.

2. Clone the repo from the official address only (copies exist; Axo checked none of them).

   git clone https://github.com/calesthio/OpenMontage.gitCopy

3. On macOS or Linux, run the setup from inside the folder. It makes a virtual environment, installs the Python and Remotion dependencies and the free Piper voice, and creates your .env file. It does not install FFmpeg: use brew install ffmpeg or sudo apt install ffmpeg.

   make setupCopy

4. On Windows, if you have no make, run this PowerShell line from inside the cloned folder (it does not install FFmpeg). If npm install fails with ERR_INVALID_ARG_TYPE, use npx --yes npm install instead.

   py -3 -m venv .venv; .\.venv\Scripts\Activate.ps1; python -m pip install -r requirements.txt; cd remotion-composer; npm install; cd ..; python -m pip install piper-tts; Copy-Item .env.example .envCopy

5. Open the folder in your coding assistant and ask in plain words, for example: "Make a 45-second animated explainer about why the sky is blue." Approve each stage when it pauses.

6. Optional: every key in .env is optional. Add the ones you have to unlock paid generators. The local board that fills in while it runs opens with:

   python -m backlot openCopy

Cheat sheet

• Repository and README.
• Think of it as a production director: your agent does the work, you approve the proposal, script, scene plan and assets. It is not hands-free.
• Free path: Piper (a local voice), free footage from Archive.org, NASA and Wikimedia Commons, plus Remotion, HyperFrames and FFmpeg. Ask for "use real footage only" for the documentary pipeline.
• Paid path: AI video and images cost per use. The maker's examples are $1.33 for a 60-second video with six clips, and about $4 and about $5 for others. The docs list Veo through fal.ai at $0.40 per second. These are the maker's numbers; Axo did not reproduce them.
• Budget: the config sets a $10 total in warn mode, which warns but does not stop at the limit; switch it to cap for a hard stop. Separately, it asks for approval on any action over $0.50 and on the first use of each paid tool.
• Vertical video: ask for the Reels, Shorts or TikTok profile (1080x1920). The default output is 1920x1080, and the repo's demo file is a 1920x1080, 30-second sci-fi trailer.
• Your coding assistant is not included and is billed on its own. Local models through Ollama or LM Studio are listed as coming soon.
• There are no GitHub releases or tags, and the repo is licensed AGPL-3.0.
• Not verified by Axo: how long it takes, the quality of a real run, or whether the setup works on your machine.

Axo's take

AXO

Haven't run it myself: this page comes from the maker's README and config files. Try it if you already work with a coding agent and want it to run a video from research to render, with you approving each gate. Start with the zero-key path: the AI generators and your agent are billed separately.