you-get

A command line tool for pulling videos, audio, and images off the web, and for streaming them straight into a media player instead of saving them.

The README describes it as a tiny command line utility for the moments when there is no other handy way to save something. GitHub's one line summary calls it a dumb downloader that scrapes the web. The pitch is about control: you watched something, you want the file, and the tool assumes you have a right to grab it. The project is written in Python, and the GitHub description leans into the scraper framing.

The capability list is short and direct. You Get downloads videos and audio from sites like YouTube, Youku, and Niconico. It streams an online video into a media player, which the README frames as no browser and no ads. It scrapes a web page to pull down images, and it can download arbitrary non HTML content such as binary files. The scraping path is called experimental and far from perfect, with the caveat that no single pattern fits every site.

There are a few ways in. The official release ships on PyPI and installs through pip with the Python 3 version. You can also download the stable or develop branches from GitHub, unzip, and put the script directory on your PATH, or clone the repository and install from there. Shell completion files for Bash, Fish, and Zsh sit in the contrib completion directory.

The info option shows available quality and formats before you commit. ffmpeg is a required dependency for joining multi part streams and for YouTube videos at 1080p or higher, and the no merge flag skips that step. Downloads pause and resume through a temporary download file, and the force flag re downloads. Output paths and file names are settable, proxy settings can be toggled or forced through proxychains, cookies load from browser files, and the player option feeds a video into mpv, vlc, or another handler.

A large supported sites table lists which sites give video, image, or audio, from YouTube and Vimeo to Tumblr, SoundCloud, Instagram, and Dailymotion, with a long set of Chinese platforms like bilibili, Youku, and 163 music in the mix. Anything not listed falls back to a universal extractor that tries to find resources on its own. The JSON output feature is explicitly flagged as not yet stabilized. The legal section makes the position clear: the authors ship the code and take no responsibility for how it is used.

For a tool this old, the README stays honest about what works and what does not. The experimental scrape path, the unstable JSON schema, and the legal disclaimer are all written down. Anyone who wants a local copy of media gets a complete manual in one file.

• Source 1 · Project repository
• Source 2 · README