Recently, I was working on a similar project and I found that grabbing the transcripts quickly leads to your IP being blocked for the transcripts.
I ended up doing the same as this person, downloading the MP4s and then transcribing myself. I was assuming it was some sort of anti LLM scraper feature they put in place.
Has anyone used this --write-auto-subs flag and not been flagged after doing 20 or so videos?
—-write-auto-subs gets your IP banned for 12/24 hours if you download video subtitles in bulk but if the subtitles are downloaded with sufficient time gap in between, the ban is not triggered.
My startup has to utilize youtube transcriptions so we just subscribe to a youtube transcriptor api hosted on rapidapi that downloads subtitles. 1$ per 1000 reqs. Pretty cheap
Unless you fetch directly from your browser. It works by getting the YouTube json including the captions track. And then you get the baseUrl to download the xml.
I wrote this webapp that uses this method: it calls Gemini in the background to polish the raw transcript and produce a much better version with punctuation and paragraphs.
It's a good call out. I leverage yt-dlp as a library for downstream tooling (archival of media to long term storage repositories), and always recommend folks rely on yt-dlp whenever possible due to the ecosystem of folks grinding to keep their extractors current. Their maintainers are both helpful and responsive.
(with that said, I do not want to diminish OP's work in any way; great job! "What I cannot build, I do not understand" - Feynman)
Minus the summarization, that is the same pipeline I use in [1] for generating listening practice Anki flashcards for foreign language students. It surprised me that nobody had really built out a program I could find around yt-dlp and Whisper for this kind of use case even a few years after it came out.
I've found the YT transcripts to be severely lacking sometimes, in accuracy and features. Especially speaker identification is really useful if you want to e.g. summarize podcasts or interviews, so if this project here delivers on that then it's definitely better than the YT transcripts.
An approach I've been using recently is to rely on pyannote/tinydiarize only for the speaker_turn timestamps, but prefer the larger model (or in this case YT's autotranscript) for the actual text.
YT transcripts definitely lack speaker ID. LLMs can infer speakers from context but miss nuance without proper speaker recognition.
I have been tackling this while building VideoToBe.com.
My current pipeline is Download Video -> Whisper Transcription with diarization -> Replace speaker tags with AI generated speaker ID + human fallback.
Reliable ML speaker identification is still surprisingly hard.
For podcast summarization, speaker ID is a game-changer vs basic YT transcripts.
I’ve had some success with running them through another LLM to have it clean up the transcription errors based on the context. But this obviously does nothing for speaker identitication.
Youtube already offers AI transcriptions on their site. As another commenter points out, you grab them with yt-dlp.
And unlike how your tool will be supported in the future, thousands of users make sure yt-dlp keeps working as google keep changing the site (currently 1459 contributors).
if you used this in earnest sufficiently, you'd know yt default transcripts are not good enough because youtube often (ok say 5% of time) fails to transcribe videos particularly livestreams and shortly after release.
the volunteer open source effort behind youtube-dl and its forks/descendants are so impressive in large part because of how many features they provide and thus have to maintain:
https://github.com/yt-dlp/yt-dlp#usage-and-options
this tool won't provide the list of available thumbnails or settings for HTTP buffer size, but I think that's a pretty reasonable tradeoff.
I tried it on a M1 Pro MBP using Docker. It's quite slow (no MPS) and there are no timestamps in the resulting transcript. But the basics are there. Truncated output:
Fetching video metadata...
Downloading from YouTube...
Generating transcript using medium model...
=== System Information ===
CPU Cores: 10
CPU Threads: 10
Memory: 15.8GB
PyTorch version: 2.7.1+cpu
PyTorch CUDA available: False
MPS available: False
MPS built: False
Falling back to CPU only
Model stored in: /home/app/.cache/whisper
Loading medium model into CPU...
100%|| 1.42G/1.42G [02:05<00:00, 12.2MiB/s]
Model loaded, transcribing...
Model size: 1457.2MB
Transcription completed in 468.70 seconds
=== Video Metadata ===
Title: 厨师长教你：“酱油炒饭”的家常做法，里面满满的小技巧，包你学会炒饭的最香做法，粒粒分明！
Channel: Chef Wang 美食作家王刚
Upload Date: 20190918
Duration: 5:41
URL: https://www.youtube.com/watch?v=1Q-5eIBfBDQ
=== Transcript ===
哈喽大家好我是王刚本期视频我跟大家分享...
Thanks for sharing. This is exactly the type of utility that vibecoding is for. It takes 5 secons to ask GPT to write a scripr to do this tailored to your specific use case. It's way faster than trying to get someone elses repo up and running.
Many channels I follow, such as Vlad Vexler, have taken measures so you can't download the transcript with yt-dlp. Furthermore, they don't provide a transcipt option on their videos. I assume this is to prevent people from just reading AI summaries, which is annoying in Vexler's case because he talks slowly and meanders around. If I really want to hear his point but don't want to listen to that then I download the video with yt-dlp and use Whisper to transcribe it.
Curious, if you don't find this "annoying", why are you still following the channel? There must be other YouTube channels that offer similar content but deliver it in a better way.
I'd be really curious to see some sort of benchmark / evaluation of these context resources against the same coding tasks. Right now, the instructions all sound so prescriptive and authoritative, yet is really hard to evaluation their effectiveness.
For better diarization quality than pyannote, check out Whisper-DiarizationX which combines Whisper with ECAPA-TDNN speaker embeddings and spectral clustering.
"The court held that merely clicking on a download button does not show consent with license terms, if those terms were not conspicuous and if it was not explicit to the consumer that clicking meant agreeing to the license."
I'm not a lawyer but I think even if you offset the legal responsibilities to the user by alerting them with copyrights prompt it's still illegal to download youtube videos.
United States v. Auernheimer, 748 F.3d 525 (3d Cir. 2014). Specifically, on page 12, footnote 5, the court states:
“We also note that in order to be guilty of accessing ‘without authorization, or in excess of authorization’ under New Jersey law, the Government needed to prove that Auernheimer or Spitler circumvented a code- or password-based barrier to access... The account slurper simply accessed the publicly facing portion of the login screen and scraped information that AT&T unintentionally published.”
Can also just fetch the subs already in YouTube rather than retranscribing. eg:
yt-dlp --write-auto-subs --skip-download "https://www.youtube.com/watch?v=7xTGNNLPyMI"
Recently, I was working on a similar project and I found that grabbing the transcripts quickly leads to your IP being blocked for the transcripts.
I ended up doing the same as this person, downloading the MP4s and then transcribing myself. I was assuming it was some sort of anti LLM scraper feature they put in place.
Has anyone used this --write-auto-subs flag and not been flagged after doing 20 or so videos?
—-write-auto-subs gets your IP banned for 12/24 hours if you download video subtitles in bulk but if the subtitles are downloaded with sufficient time gap in between, the ban is not triggered.
My startup has to utilize youtube transcriptions so we just subscribe to a youtube transcriptor api hosted on rapidapi that downloads subtitles. 1$ per 1000 reqs. Pretty cheap
Yep, this happened to me & got IP banned for a day.
Unless you fetch directly from your browser. It works by getting the YouTube json including the captions track. And then you get the baseUrl to download the xml.
I wrote this webapp that uses this method: it calls Gemini in the background to polish the raw transcript and produce a much better version with punctuation and paragraphs.
https://www.appblit.com/scribe
Open source with code to see how to fetch from YouTube servers from the browser https://ldenoue.github.io/readabletranscripts/
It's a good call out. I leverage yt-dlp as a library for downstream tooling (archival of media to long term storage repositories), and always recommend folks rely on yt-dlp whenever possible due to the ecosystem of folks grinding to keep their extractors current. Their maintainers are both helpful and responsive.
(with that said, I do not want to diminish OP's work in any way; great job! "What I cannot build, I do not understand" - Feynman)
Same, yup. OP is indeed already using yt-dlp for the video download. (Then Whisper for transcribing, Ollama/lmstudio/OpenAI for summarizing)
Minus the summarization, that is the same pipeline I use in [1] for generating listening practice Anki flashcards for foreign language students. It surprised me that nobody had really built out a program I could find around yt-dlp and Whisper for this kind of use case even a few years after it came out.
[1]: https://github.com/hiAndrewQuinn/audio2anki
I've found the YT transcripts to be severely lacking sometimes, in accuracy and features. Especially speaker identification is really useful if you want to e.g. summarize podcasts or interviews, so if this project here delivers on that then it's definitely better than the YT transcripts.
An approach I've been using recently is to rely on pyannote/tinydiarize only for the speaker_turn timestamps, but prefer the larger model (or in this case YT's autotranscript) for the actual text.
Check out https://ldenoue.github.io/readabletranscripts/ and the website https://www.appblit.com/scribe that use Gemini to post correct the raw transcripts
YT transcripts definitely lack speaker ID. LLMs can infer speakers from context but miss nuance without proper speaker recognition.
I have been tackling this while building VideoToBe.com. My current pipeline is Download Video -> Whisper Transcription with diarization -> Replace speaker tags with AI generated speaker ID + human fallback.
Reliable ML speaker identification is still surprisingly hard. For podcast summarization, speaker ID is a game-changer vs basic YT transcripts.
I’ve had some success with running them through another LLM to have it clean up the transcription errors based on the context. But this obviously does nothing for speaker identitication.
IIRC YT also has a "private" API you can call directly (or via an npm package: youtube-transcribe).
(I'm using it in https://butter.sonnet.io)
Yep. You can also automatically save them if you use mpv to watch YT: https://github.com/nick-s-b/mpv-transcript discovered this script yesterday.
[dead]
For (English only) speech-to-text, NVIDIA's Parakeet-V2 is significantly faster than Whisper and I found it to be more accurate.
https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2
For Apple Silicon (MLX) https://huggingface.co/senstella/parakeet-tdt-0.6b-v2-mlx
Compared to all Whister models? Or the faster ones? And which version of Whisper? All for a faster, more accurate model, but need a bit more.
All of them, in my experience.
Fair, looking at the ASR leaderboards it is truly better - https://huggingface.co/spaces/hf-audio/open_asr_leaderboard and NVIDIA's Canary might be even better? Will try these out. Appreciate bringing these to my attention!
Youtube already offers AI transcriptions on their site. As another commenter points out, you grab them with yt-dlp.
And unlike how your tool will be supported in the future, thousands of users make sure yt-dlp keeps working as google keep changing the site (currently 1459 contributors).
if you used this in earnest sufficiently, you'd know yt default transcripts are not good enough because youtube often (ok say 5% of time) fails to transcribe videos particularly livestreams and shortly after release.
youtube also blocks transcript exports for some things like https://youtubetranscript.com/
retranscribing is necessary and important part of the creator toolset.
the volunteer open source effort behind youtube-dl and its forks/descendants are so impressive in large part because of how many features they provide and thus have to maintain: https://github.com/yt-dlp/yt-dlp#usage-and-options this tool won't provide the list of available thumbnails or settings for HTTP buffer size, but I think that's a pretty reasonable tradeoff.
Hey all, I built a 100% free (no-signup) youtube summarizer: "https://youtube-summarizer-lime.vercel.app/". Accurate summaries in under 8 seconds.
How did you get around youtube blocking cloud IP ranges? Are you suing residential proxies?
bookmarked, thanks, the top google search results always require sign-up. frustrating state of the internet
I made a tool like this a while ago which was useful for transcribing a whole playlist automatically using whisper:
https://github.com/Dicklesworthstone/bulk_transcribe_youtube...
I ended up turning a beefed up version of it which makes polished written documents from the raw transcript, you can try it at
https://youtubetranscriptoptimizer.com/
Two similar Show HN projects:
- This python one is more amenable to modding into your own custom tool: https://hw.leftium.com/#/item/44353447
- Another bash script: https://hw.leftium.com/#/item/41473379
---
They all seem to be built on top of:
- yt-dlp to download video
- whisper for transcription
- ffmpeg for audio/video extraction/processing
I tried it on a M1 Pro MBP using Docker. It's quite slow (no MPS) and there are no timestamps in the resulting transcript. But the basics are there. Truncated output:
> Falling back to CPU only
Patient: “Doctor, it hurts when I do this.”
Doctor: “don’t do that”
I mean yeah, but also
Doctor: do this
Patient: I tried doing this and it's not good
Doctor: actually you need a device for $5000 lol
I’ve been using this free tool. It gives quality diarized transcripts https://contentflow.megalabs.co
Did you build this? I'm looking for an API that does this.
I built this. The API is on the way! You can sign up for updates here: https://contentflow.megalabs.co/api-interest
Thanks for sharing. This is exactly the type of utility that vibecoding is for. It takes 5 secons to ask GPT to write a scripr to do this tailored to your specific use case. It's way faster than trying to get someone elses repo up and running.
Selfware.
https://old.reddit.com/r/ChatGPTCoding/comments/1lusr07/self...
Gonna be lots of posts of selfware like that soon.
I like it, though I'm sure we'll end up being stuck with "vibe ware"
I think you either coined (kudos) or spotted the true "term du jour" here.-
people don't even get it :-]
Sure thing ...
And, yes, indeed, AI-coding is order-of-magnitude having an effect along the lines that "low-code" was treading ...
... also, for less-capable coders or "borderline" coders the effort/benefit equation has radically shifted.-
Many channels I follow, such as Vlad Vexler, have taken measures so you can't download the transcript with yt-dlp. Furthermore, they don't provide a transcipt option on their videos. I assume this is to prevent people from just reading AI summaries, which is annoying in Vexler's case because he talks slowly and meanders around. If I really want to hear his point but don't want to listen to that then I download the video with yt-dlp and use Whisper to transcribe it.
Curious, if you don't find this "annoying", why are you still following the channel? There must be other YouTube channels that offer similar content but deliver it in a better way.
Vlad is a smart guy but slow. Think of him as a brilliant snail.
... the ... slower ... the guy the ... less ... content ... and ... more ... advertising.-
I did something similar piping the output of the youtube-transcript-api python package to openAI's api: https://github.com/DavidZirinsky/tl-dw/
Always fascinated to read CLAUDE.md files that are appearing in more and more open source projects: https://github.com/pmarreck/yt-transcriber/blob/yolo/CLAUDE....
I'd be really curious to see some sort of benchmark / evaluation of these context resources against the same coding tasks. Right now, the instructions all sound so prescriptive and authoritative, yet is really hard to evaluation their effectiveness.
Interesting project! I've been working on a project in this space myself (WaveMemo)
I must say, speaker diarization is surprisingly tricky to do. The most common approach seems to be to use pyannote, but the quality is not amazing...
For better diarization quality than pyannote, check out Whisper-DiarizationX which combines Whisper with ECAPA-TDNN speaker embeddings and spectral clustering.
Youtube's T&C don't allow downloading youtube audio/video. How do other services get away with it?
"The court held that merely clicking on a download button does not show consent with license terms, if those terms were not conspicuous and if it was not explicit to the consumer that clicking meant agreeing to the license."
https://en.m.wikipedia.org/wiki/Specht_v._Netscape_Communica...
I'm not a lawyer but I think even if you offset the legal responsibilities to the user by alerting them with copyrights prompt it's still illegal to download youtube videos.
United States v. Auernheimer, 748 F.3d 525 (3d Cir. 2014). Specifically, on page 12, footnote 5, the court states:
“We also note that in order to be guilty of accessing ‘without authorization, or in excess of authorization’ under New Jersey law, the Government needed to prove that Auernheimer or Spitler circumvented a code- or password-based barrier to access... The account slurper simply accessed the publicly facing portion of the login screen and scraped information that AT&T unintentionally published.”
I think they use rotating IP/Proxy services
Might be, but I think google would still be able to chase them down.
On this note, is Ytube also the best transcriber of foreign languages or is there something better?
Will this make Google mad at me and cancel/freeze all my Google services ?
I vibecoded something similar for myself, transcribes and summarizes the content into article format: https://github.com/senko/scribe
Uses yt-dlp, whisper, and a LLM (Gemini hardcoded because it handles long contexts well, but easy to switch) for summarizer.
I dislike podcast as a format (S/N level way too low for my taste), so use this whenever I want to get a tldr of some episode.
I should check out the SOTA models and improve the summarization prompt, but aren't in a hurry as this works pretty well for my needs already.
[dead]