YouTube innertube class library for node-js; session, searches, channels, playlists, videos, comments and downloads.

• Simple, efficient, fast, powerful.
• No Google developer key required.
• The innertube api is what the YouTube website itself uses to efficiently deliver search, channel and video information (json only).
• The downloader is a forked process allowing for concurrent non-blocking, high quality downloads.

• Session - manage your Youtube session/player - deciphering, encoding and hashing - enables seemless search, channel, playlist, video and download requests.
• Search - fetch videos, playlists and channels from specific search requests.
• Channel - fetch metadata, videos, playlists, associated channels or search specific channels.
• Playlist - fetch videos from specific playlists
• Video - fetch metadata and stream information deciphered/encoded to avoid throttling - ensure reliable and fast downloads.
• Comment - helper class for the download and management of video comments and threads.
• Download - a convenience class for easy once-off, sessionless, downloads.

See the wiki for greater detail.

const ytcog = require('ytcog');
await ytcog.dl(videoOptions[, cookie, userAgent, proxy, debug]);

videoOptions (object) See the wiki for all videoOptions.

cookie (string) is optional. With a cookie, everything will work. Without it, age-restricted video streams will not be retrieved and there might be some rate-limiting (although none reported so far)

userAgent (string) is optional. Since ytcog emulates a browser session, you can make all requests use your browser's user agent. If you supply a userAgent, you must supply a cookie, even if it is an empty string.

proxy (string) is optional. Provide a proxy agent string for all session https requests, i.e:
await ytcog.dl({id:'5qwDrjTinMk'},'','','http://127.0.0.1:8000');

debug (boolean) if true debug information is sent to the console.

NB: If you are downloading multiple videos (i.e. from search results, playlists or channels) then maintianing a session and using video.download() is much more efficient than running ytcog.dl() on each video.

const ytcog = require('ytcog');
const session = new ytcog.Session([cookie, userAgent, proxy]);
await session.fetch();
console.log(`session status: ${session.status}`);

cookie and userAgent are optional, but in order to obtain them log onto YouTube in your browser. Goto settings > ... > developer tools. Refresh the page. Goto network>headers. Find the "www.youtube.com" entry. In the request headers you will find "cookie" and "user-agent". Pass these string values into your ytcog sessions. If you supply a userAgent, you must supply a cookie, even if it is an empty string.

proxy (string) is optional. Provide a proxt agent string for all session https requests, i.e:
let session = new ytcog.session('','','http://127.0.0.1:8000')

A session object is required to create search, channel, playlist and video objects.

const search = new ytcog.Search(session, searchOptions);
await search.fetch();

session (Object) the session object, see above.

searchOptions (Object) See the wiki for all search options.

Search again with different options:

await search.fetch({items: 'videos', period:'year', order: 'views', features: 'hd', quantity: 500 });

Examine the results in an array of Video objects:

search.videos.forEach((video)=>{
    // do something with the results, like collect and display their streams
    await video.fetch();
    console.log(video.info());
    console.log(video.streamInfo);
});

Also search for playlists, channels and movies that match your search term

await search.fetch({items:'playlists'});
await search.fetch({items:'channels'});
await search.fetch({items:'movies'});

Iterate through the current results with:

search.results.forEach((item)=>{});

Or the accumulated results

search.playlists.forEach((playlist)=>{...});
search.channels.forEach((channel)=>{...});
search.videos.forEach((video)=>{...});

const channel = new ytcog.Channel(session, channelOptions);
await channel.fetch();

channelOptions See wiki for all channel options.

Get channel playlists

await channel.fetch({items: 'playlists', order: 'updated', quantity: 90});

Get associated channels

await channel.fetch({items: 'channels'});

Search a channel

await channel.fetch({items: 'search', query: 'vlogs'});

Iterate through the results with:

channel.results.forEach((item)=>{});  //current
channel.videos.forEach((video)=>{...}); //accumulated
channel.playlists.forEach((playlist)=>{...}); //accumulated
channel.channels.forEach((chan)=>{...}); //accumulated

const playlist = new ytcog.Playlist(session, playlistOptions);
await playlist.fetch();

playlistOptions See wiki for all playlist options.

Get 100 videos from a playlist

await playlist.fetch({quantity:100});

Get all the videos from a playlist

await playlist.fetch({quantity: playlist.videoCount});

Iterate through the results with:

playlist.results.forEach((video)=>{...}) //current
playlist.videos.forEach((video)=>{...}); //accumulated

Get metadata, media and stream information:

const video = new ytcog.Video(session, videoOptions);
await video.fetch();

Get comments

const video = new ytcog.Video(session, videoOptions);
await video.fetchComments(commentOptions);

Or just download:

const video = new ytcog.Video(session, videoOptions);
await video.download();

videoOptions See wiki for all video options.
commentOptions See wiki for comment options.

Check the examples folder for more clarity on usage of Session, Search, Channel, Playlist and Video classes.

To run the examples:

ytcog> node examples/session_test
ytcog> node examples/search_test [query]
ytcog> node examples/channel_test [id]
ytcog> node examples/playlist_test [id]
ytcog> node examples/video_test [id]
ytcog> node examples/dl_test [id]

npm install ytcog

YouTube can and will change how their innertube api works at any time. So potential disruptions are likely in the future. I will try to evolve and adapt this library asap, but without gaurantees.

Try out the command line interface (CLI) to this library:

• ytcog-dl

To the following node-js projects on which ytcog has a dependency:

• miniget (robust web requests)
• vm2 (bullet-proof execution of 3rd party javascript)
• ffmpeg-static (muxing video and audio downloads; embedding metadata as necessary)
• sanitize-filename (as the name suggests)
• proxy-agent (enables requests via a proxy)