Clip Pipe

End-to-end automation pipeline for turning long videos into short-form content and uploading to YouTube + Telegram automatically.

This project handles downloading videos, clipping by timestamp, formatting for shorts, adding subtitles, and sending the result to YouTube and Telegram.

Built for batch processing and unattended execution.

---

Features

• Download videos from YouTube by URL
• Clip videos using start and end timestamps
• Auto-generate short-form vertical videos
• Subtitle embedding (auto)
• Video cropping and positioning
• Watermark based account
• Brainrot
• Multi-account support
• Upload to YouTube
• Send video to Telegram (title + description)
• Upload scheduling
• Local file support
• Proxy support
• Web Manager UI for job management
• Fully JSON-driven workflow

---

Installation (Nix)

This project uses Nix shell for dependency management.

All required system tools and Python packages are provided automatically.

Requirements

• Linux or WSL2 (Windows, not tested yet)
• Nix package manager

---

Install Nix

Windows (WSL2)

wsl --install

curl -L https://nixos.org/nix/install | sh

Linux

curl -L https://nixos.org/nix/install | sh

Linux (Nixos)

nix-shell

Env Setup

# Telegram bot for sending videos
# create bot using this bot official telegram @BotFather
TELEGRAM_TOKEN=
TELEGRAM_CHAT_ID=

# (optional)
# Comma-separated list of proxy URLs
PROXIES=proxy1,proxy2

# (optional) Delay between jobs (in seconds)
MIN_DELAY=30
MAX_DELAY=60

# (optional) How often to check for pending jobs (in seconds)
CHECK_INTERVAL=30

# (optional) Max retries on failure
MAX_RETRIES=3

Authentication Setup

Google (YouTube)

Follow /src/auth/GOOGLE_OAUTH.md which covers:

• Google Cloud project creation
• OAuth consent screen
• Client ID setup
• Refresh token generation

Account Configuration

Create an accounts/ folder in project root, then create a folder for your account name (e.g., obrolan_clip/). Inside that folder, add client_secret.json.

accounts/
└── obrolan_clip/
     └── client_secret.json  // for YouTube
└── other_account/
     └── client_secret.json

Full JSON Configuration

[
  {
    "date": "2026-01-28,12:00", # date for scheduling 
    "status": "pending",        # if already executed will change to completed
    "items": [
      {
        "url": "",           # url youtube (or use local for local file)
        "local": "",         # local file path (alternative to url)
        "start": "00:06:18", # time start clip
        "end": "00:07:18",   # end  time clip (result will be 01:00 minute shorts video)
        "crop": null,        # crop position: "left", "center", "right" or null for no crop  
        "caption": false,    # if set to false will skip fast-whisper (auto generate subtitle)
        "model": "small",    # whisper model: tiny, small, medium, large
        "brainrot": false,   # if true it will added other video below the original shorts
        "tests": false,      # if true will skip upload to YouTube and Telegram, save to media/shorts/ 
        "account": "other_username", # account name based on folder inside accounts/
        "title": "",         # title video
        "description": ""    # description (pass tags is accepted)
      }
    ]
  }
]

Web Manager (Optional)

For a visual interface to manage jobs, use the Web Manager:

python3 start-web-manager.py

Then open http://localhost:9876 in your browser.

---

Run

Enter folder ./src then run

python3 job_runner.py

Disclaimer

This project is intended for educational and automation purposes only.

You are responsible for:

• API usage
• Platform rate limits
• Content ownership
• Compliance with YouTube and Telegram policies

Buy Me a Lambo

SOL: E1D96h652zzTzh8QyT1Dq8AqECT3j7dbJxJMiumnBV7c

EVM: 0x965244d54FFdf69A287e122090d320ae648f4cf8

TODO

• Auto detect most replay (get timestamp)