Skip to content
Source preview coming soon. Join the waitlist for the opening.
Cutgraph
Get started

Your first cut, a few installs away.

The developer preview is about to open. Join the waitlist for the source release; the requirements and install steps below preview what is coming.

Join the waitlist

Grab a coffee. I'll wait.

Requirements

What your machine needs.

Operating system

  • Windows 11 with an NVIDIA GPU (tested) Shipped
  • Linux (Ubuntu in CI), best-effort Preview
  • macOS Not supported

Software

  • Python 3.12 or later
  • Node 24 and npm
  • ffmpeg and ffprobe with libx264, on your PATH
  • A current browser (Studio opens at a local address on your machine)

Graphics card

Not required for editing, preview or CPU export. An NVIDIA GPU makes export faster through NVENC and transcription faster through CUDA. Without one, both fall back to the CPU.

Disk and network

Room for your media, plus about 3.5 GB for the speech models if you add transcription. They download once on first use. Your audio never leaves your machine; to stop model-host checks as well, provision the cache and turn on Hugging Face offline mode.

The install, at a glance

Four steps, then roll.

A preview of the Windows source install, for when the repository opens. These steps need the source release and are not available to run yet.

  1. STEP 01

    Get the code

    When the preview opens, clone the repository or download its source release. Join the waitlist above to hear when it is ready.

  2. STEP 02

    Install the core

    A Python virtual environment with hash-locked dependencies, then the UI and motion engine packages:

    python -m venv .venv
.venv\Scripts\python -m pip install --require-hashes -r requirements-core.txt
npm --prefix ui ci
npm --prefix motion-core ci

  3. STEP 03

    Bootstrap and build

    Run the bootstrap script, then build the interface once:

    .venv\Scripts\python scripts/bootstrap.py
npm --prefix ui run build

  4. STEP 04

    Start Studio

    Start the server from the virtual environment, then open http://127.0.0.1:8765/editor/ in your browser. It only listens on your own machine. Optional keys go in Settings.

    .venv\Scripts\python scripts/studio_app.py --port 8765
Optional add-ons

Add what you need, skip what you don't.

Transcription

Shipped

The local WhisperX pack. English only. Uses CUDA when available, CPU otherwise. Needed to edit footage by its words or to align narration.

Co-pilot

Shipped

The optional co-pilot add-on plus your own Anthropic API key, entered in Settings or as ANTHROPIC_API_KEY. Billed by Anthropic.

Stock search

Shipped

Free Pexels and Pixabay API keys let the drawer search stock B-roll. Imports keep their attribution.

Stuck?

The issue tracker and discussions open with the source preview. For now, join the waitlist or contact us from the Community page. Support is best-effort while we're in preview. How to reach us

Get started with the developer preview | Cutgraph Studio