Chrome extension adjusts video speed based on how fast the speaker is talking (opens in new tab)

This is a very clever idea. Unfortunately I dislike installing extensions, because so many of them seem to get taken over by nefarious owners. Sadly it's poisoned the whole extension platform for me.

FYI: On Youtube, the keyboard shortcut for changing the video speed is simply pressing < or >

Very interesting idea, but doesn't seem to work very well. I tried on three videos. On two it miscounted to an absurdly low rate of ~1 syllable/s and played it at the max speed which was illegible and in one it failed to detect any speech at all.

For most videos the speaker talks in a relatively constant speed throughout the video. I've been using Video Speed Controller on Firefox for years. Sadly, it's not actively monitored by Mozilla for security issues even though it has 100k+ downloads. It basically lets you change the speed by whatever increments you want with a key. If the speaker slows down, tap-tap-tap, you can speed it up. I don't really see the need for algorithms or AI to solve what's already almost-solved.

I don't get why this addon isn't included in Firefox to begin with. Maybe because it doesn't just work only on standard HTML5 video files, but has support for specific sites and their players, too? But I've seen several updates on Firefox (not about the addon) that are along the lines of "solved perf issues on $site". So it's not like Firefox is trying to be site-agnostic.

This would be a killer feature for a podcast app. I can't stomach NPR content unless it's sped up to almost 2x, with the exception of a few fast talking hosts.

I have a policy against installing extensions but I might make an exception for this one. I default to 2x and often have to go to 3x if the speaker is really slow. The one exception is music production related videos and it is sometimes excruciating to get through. This is most likely because I'm usually watching some kind of instructional or educational video and just want the content and have no interest in being "entertained". That said, I understand that you have to game YouTube's algo and rules in order to make even a tiny bit of money. Imagine if every channel cut the first 30 seconds of every video intro and everyone could just get to the point.

Damn it!

Now I am called The Snail Speaker. :(

What I found interesting is ... if you have some recorded lectures from a university, some speakers are ultra-fast at 1x speed; and some others are acceptable once at 1.5 or 1.75.

Most I find difficult to understand at 2.0x though, so I mostly settled between the range of 1.25 to 1.5, sometimes 1.75 (depends on the topic too, of course; some topic is harder to understand, but here my thought was that some are able to speak fast and others can not. It's almost a skill to have or not have).

I can thing of a couple of content creators I've encountered who make fast voice presentations almost impossible to follow. almost as if that's the point -- like that academic debating style where speakers just speed-recite as many arguments as they can -- like the content is actually meant to be slower and their goal is data compression or something ...

This is really cool. I build Chrome extensions too and the hardest part with anything audio related is getting the detection right across different types of content. Lectures vs podcasts vs casual vlogs all have very different speech patterns. I wonder if this could also work with non-English content since syllable detection probably breaks down with tonal languages.

Damn, I was working on exactly the same idea not too long ago. It never really went anywhere because I couldn't find a speech rate detection algorithm that worked well enough to make it practically useful. I hope some more mathematically inclined people take a look at this idea and find a way to make it work.

Pretty cool, would be great if it could detect accented speech and have a different normalized speed for that. I listen to most people at 2x or more, but I can't usually understand accented English quite that fast.

How good is the web's support for having subtitles? I'd love a version of this that just worked off of subtitles, rather than audio processing.

Wow. I’ve been looking for this from a long time.

I would never install an extension with a 1000 foot pole, because of its seriously flawed security model. I think in 2026 if you really badly need an extension as an end user, write your own or pay someone to do it. I would also accept open source ones, since I can review the code.

Your extension can and will likely read all your data and likely has broad permissions to just fax all your Gmail to remote cloud server.