Skip to Content
shardreel

Getting started

Clone the open-source repo, install the requirements, build shardreel and run your first job on your Mac.

Apple Silicon only

shardreel targets M-series Macs. It relies on Metal for Whisper and VideoToolbox for encoding.

shardreel is open source. You run it from a local clone of the repository: github.com/shardreel/clipforge.

Requirements

ToolWhy it is needed
Rust toolchainBuilds the backend
bunBuilds the web UI
cmakeBuilds whisper.cpp, which the Whisper bindings compile
ffmpeg with libassDecoding, filters, encoding, and burned-in captions
yt-dlpDownloading videos from URLs

Install the command-line tools

brew install cmake yt-dlp ffmpeg-full
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
curl -fsSL https://bun.sh/install | bash

ffmpeg-full is a keg-only build that includes libass. It does not replace an ffmpeg you already have. shardreel looks for it in /opt/homebrew/opt/ffmpeg-full/bin first and falls back to your PATH.

Clone the repository

git clone https://github.com/shardreel/clipforge.git
cd clipforge

To pin a release instead of the latest main, run git tag and git checkout <tag>.

Build the web UI

cd web
bun install
bun run build

Run the app

cargo run --release

The first build compiles whisper.cpp and takes a few minutes. When it is ready, the server prints its address:

clipforge listening on http://127.0.0.1:8787

Check the setup

Open http://127.0.0.1:8787 and go to Settings. The System card lists each requirement as OK or Missing, with the command that fixes it.

Run your first job

  1. Open New job.
  2. Drop in a video, or switch to Path or URL and paste a path or a YouTube or Twitch link.
  3. Leave the defaults and press Generate clips.
  4. Watch the stages on the job page. When it finishes, play the clips in the grid.

The first job with speech recognition downloads the Whisper model, about 1.6 GB for the default large-v3-turbo, into ~/.clipforge/models. Later runs reuse it.

Clips are saved to ~/Movies/clipforge/ by default, in a folder named after the video. Change the output folder per job.

Update to the latest version

cd clipforge
git pull
cd web && bun install && bun run build && cd ..
cargo run --release

Your settings, models and cache live in ~/.clipforge, so updating does not touch them.

Develop the UI

Run the backend and the Vite dev server side by side:

cargo run --release
cd web && bun run dev

Open http://localhost:5173. Vite proxies /api to the backend on port 8787.

Contribute

Found a bug or want a feature? Open an issue or send a pull request. Fork the repo, create a branch, run cargo run --release and cd web && bun run dev to test your change, then open a PR against main. If something fails during setup, check Troubleshooting first.

Next steps

Edit on GitHubSpotted a mistake? Suggest a change and open a pull request.

On this page