Getting Started¶
There are three ways to run ClipPulse AI.
1. Local (recommended for development)¶
Requirements: Python 3.12+, FFmpeg, Deno, a free Gemini API key.
git clone https://github.com/YusifMurad/ClipPulse-AI.git
cd ClipPulse-AI
chmod +x start.sh
./start.sh
Open http://localhost:5555. On first run a Python virtualenv is created for you.
Manual setup¶
cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
export PATH="$HOME/.deno/bin:$PATH" # Deno for yt-dlp
sudo apt install ffmpeg # or: brew install ffmpeg
python3 server.py
2. Docker¶
docker compose up -d --build
# → http://localhost:5555
Or pull the prebuilt image:
docker run -p 5555:5555 -v "$PWD/output:/app/output" \
ghcr.io/yusifmurad/clippulse-ai:latest
3. GitHub Codespaces¶
Click Open in Codespaces on the repo, wait ~2 minutes, then open the
forwarded port 5555 in your browser.
First run¶
- Settings → paste your Gemini API key (stored server-side, never sent to the client).
- Paste a YouTube URL or upload an MP4.
- Click Create Clips.
- Preview, edit, and download.