The developers of the yt-dlp project, which continues the development of the youtube-dl tool for downloading audio and video from online services, have announced that starting with today's release of yt-dlp 2025.11.12, a separate JavaScript engine is required for fully downloading content from YouTube. The use of a JavaScript engine is explained by the introduction of checks on YouTube, which require the execution of code in JavaScript to pass.
Support for working with YouTube without JavaScript has been declared deprecated in yt-dlp. It is noted that some videos may still be downloadable without using the JavaScript engine, but the number of supported formats is limited, and it is expected that the situation will worsen over time. Failures occurring when downloading content from YouTube without the application of JavaScript will now be regarded by the yt-dlp developers not as errors, but as issues that are not subject to correction. It is expected that working with YouTube without JavaScript will become impossible in the future.
The supported JavaScript runtimes listed include Deno (recommended), Node.js, QuickJS, QuickJS-ng, and Bun. Additionally, for the JavaScript runtime to work in yt-dlp, the yt-dlp-ejs component must be installed, which is included in the official builds. It is sufficient to install the desired JavaScript runtime through the distribution's package manager, after which it can be activated using the option "--js-runtimes [deno|quickjs|bun|node]". The Deno engine is activated by default, so the "--js-runtimes" option can be omitted for it. When manually installing the JavaScript runtime, the path to the engine can be passed in the format "--js-runtimes deno:/path/to/deno".
Among the changes unrelated to YouTube in the new version, the addition of new content extractors for services 1tv.ru/live, classics.nascar.com, mux.com, and ardaudiothek.de stands out. The extractor for ntv.ru has been overhauled. The performance of extractors for twitch, play.tv, googledrive, bunnycdn, kika, nbc, and webarchive has been improved.
Source: opennet.ru
