<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: BulkTranscripts youtube </title>
    <description>The latest articles on DEV Community by BulkTranscripts youtube  (@rapidshortsai_012681b0e02).</description>
    <link>https://dev.to/rapidshortsai_012681b0e02</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4138049%2F62b3f627-58cc-4e26-8769-aa792058fdf0.png</url>
      <title>DEV Community: BulkTranscripts youtube </title>
      <link>https://dev.to/rapidshortsai_012681b0e02</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/rapidshortsai_012681b0e02"/>
    <language>en</language>
    <item>
      <title>yt-dlp subtitles: the commands that work, and 3 traps that cost me an afternoon</title>
      <dc:creator>BulkTranscripts youtube </dc:creator>
      <pubDate>Wed, 04 Nov 2026 18:29:31 +0000</pubDate>
      <link>https://dev.to/rapidshortsai_012681b0e02/yt-dlp-subtitles-the-commands-that-work-and-3-traps-that-cost-me-an-afternoon-18m6</link>
      <guid>https://dev.to/rapidshortsai_012681b0e02/yt-dlp-subtitles-the-commands-that-work-and-3-traps-that-cost-me-an-afternoon-18m6</guid>
      <description>&lt;p&gt;If you only want the subtitles from a YouTube video, yt-dlp can do it without downloading the video. The flags look simple. Then you run them on a real channel and get a 429, a folder full of junk languages, or a re-run that starts again from zero.&lt;/p&gt;

&lt;p&gt;Everything below was run with yt-dlp 2026.07.04 on real videos this week.&lt;/p&gt;

&lt;h2&gt;
  
  
  The one command most people want
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;yt-dlp &lt;span class="nt"&gt;--skip-download&lt;/span&gt; &lt;span class="nt"&gt;--write-subs&lt;/span&gt; &lt;span class="nt"&gt;--write-auto-subs&lt;/span&gt; &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;--sub-langs&lt;/span&gt; en &lt;span class="nt"&gt;--sub-format&lt;/span&gt; srt &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;-o&lt;/span&gt; &lt;span class="s2"&gt;"%(id)s.%(ext)s"&lt;/span&gt; &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="s2"&gt;"https://www.youtube.com/watch?v=dQw4w9WgXcQ"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;What each part does:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;--skip-download&lt;/code&gt; keeps the video off your disk. You only get the subtitle file.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;--write-subs&lt;/code&gt; takes subtitles the creator uploaded.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;--write-auto-subs&lt;/code&gt; takes YouTube's auto-generated captions.&lt;/li&gt;
&lt;li&gt;With both set, yt-dlp prefers the creator's track and falls back to the auto one. On the video above it picked the uploaded track, which has proper punctuation and the ♪ marks.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;--sub-format srt&lt;/code&gt; asks YouTube for SRT directly, so no ffmpeg step is needed.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;To see what a video has before downloading anything:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;yt-dlp &lt;span class="nt"&gt;--list-subs&lt;/span&gt; &lt;span class="s2"&gt;"https://www.youtube.com/watch?v=dQw4w9WgXcQ"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Trap 1: &lt;code&gt;--sub-langs "en.*"&lt;/code&gt; gets you a 429
&lt;/h2&gt;

&lt;p&gt;The yt-dlp help text uses &lt;code&gt;en.*&lt;/code&gt; as an example, so it's the first thing everyone tries. On a popular video, &lt;code&gt;--list-subs&lt;/code&gt; shows why it's a bad idea: the automatic captions include dozens of machine-translated tracks, and the codes look like this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;en-en      English from English
ab-en      Abkhazian from English
en-de-DE   ...
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;code&gt;en.*&lt;/code&gt; matches &lt;code&gt;en&lt;/code&gt;, &lt;code&gt;en-orig&lt;/code&gt;, &lt;code&gt;en-en&lt;/code&gt;, &lt;code&gt;en-de-DE&lt;/code&gt; and more. yt-dlp requests every one of them back to back, and YouTube answered my test with:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;ERROR: Unable to download video subtitles for 'en-de-DE': HTTP Error 429: Too Many Requests
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The fix is to ask for exactly what you want:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="nt"&gt;--sub-langs&lt;/span&gt; en            &lt;span class="c"&gt;# just English&lt;/span&gt;
&lt;span class="nt"&gt;--sub-langs&lt;/span&gt; &lt;span class="s2"&gt;"en,es"&lt;/span&gt;       &lt;span class="c"&gt;# English and Spanish&lt;/span&gt;
&lt;span class="nt"&gt;--sub-langs&lt;/span&gt; &lt;span class="s2"&gt;"en,-live_chat"&lt;/span&gt;  &lt;span class="c"&gt;# English, never the live-chat replay&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Trap 2: auto-caption VTT files are not readable text
&lt;/h2&gt;

&lt;p&gt;If you grab auto captions as VTT and open the file, you see this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;00:00:21.800 --&amp;gt; 00:00:25.950 align:start position:0%
We're no strangers to
love.&amp;lt;00:00:22.800&amp;gt;&amp;lt;c&amp;gt; You&amp;lt;/c&amp;gt;&amp;lt;00:00:23.039&amp;gt;&amp;lt;c&amp;gt; know&amp;lt;/c&amp;gt;&amp;lt;00:00:23.279&amp;gt;&amp;lt;c&amp;gt; the&amp;lt;/c&amp;gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Each word has its own timing tag, and every line appears twice, because auto captions "roll" on screen. Feed that to an LLM and you pay for every word twice.&lt;/p&gt;

&lt;p&gt;This turns it into clean text:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;re&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;sys&lt;/span&gt;

&lt;span class="k"&gt;def&lt;/span&gt; &lt;span class="nf"&gt;vtt_to_text&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;path&lt;/span&gt;&lt;span class="p"&gt;):&lt;/span&gt;
    &lt;span class="n"&gt;lines&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;[]&lt;/span&gt;
    &lt;span class="k"&gt;for&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="ow"&gt;in&lt;/span&gt; &lt;span class="nf"&gt;open&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;path&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;encoding&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;utf-8&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;):&lt;/span&gt;
        &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;strip&lt;/span&gt;&lt;span class="p"&gt;()&lt;/span&gt;
        &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="ow"&gt;not&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="ow"&gt;or&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;--&amp;gt;&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt; &lt;span class="ow"&gt;in&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="ow"&gt;or&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;startswith&lt;/span&gt;&lt;span class="p"&gt;((&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;WEBVTT&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;Kind:&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;Language:&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)):&lt;/span&gt;
            &lt;span class="k"&gt;continue&lt;/span&gt;
        &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;re&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;sub&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sa"&gt;r&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;&amp;lt;[^&amp;gt;]+&amp;gt;&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="sh"&gt;""&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;  &lt;span class="c1"&gt;# drop the per-word timing tags
&lt;/span&gt;        &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="ow"&gt;and&lt;/span&gt; &lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="ow"&gt;not&lt;/span&gt; &lt;span class="n"&gt;lines&lt;/span&gt; &lt;span class="ow"&gt;or&lt;/span&gt; &lt;span class="n"&gt;line&lt;/span&gt; &lt;span class="o"&gt;!=&lt;/span&gt; &lt;span class="n"&gt;lines&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="mi"&gt;1&lt;/span&gt;&lt;span class="p"&gt;]):&lt;/span&gt;  &lt;span class="c1"&gt;# auto captions repeat each line
&lt;/span&gt;            &lt;span class="n"&gt;lines&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;append&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;line&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
    &lt;span class="k"&gt;return&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt; &lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;join&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;lines&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;

&lt;span class="nf"&gt;print&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="nf"&gt;vtt_to_text&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;sys&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;argv&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="mi"&gt;1&lt;/span&gt;&lt;span class="p"&gt;]))&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Output:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;[Music] We're no strangers to love. You know the rules and so do I. ...
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;If you want timestamps, ask for &lt;code&gt;--sub-format srt&lt;/code&gt; instead. The SRT version of the same captions doesn't have the rolling duplicates.&lt;/p&gt;

&lt;h2&gt;
  
  
  Trap 3: &lt;code&gt;--download-archive&lt;/code&gt; does nothing with &lt;code&gt;--skip-download&lt;/code&gt;
&lt;/h2&gt;

&lt;p&gt;For a whole channel you want re-runs to skip what you already have, so you add &lt;code&gt;--download-archive done.txt&lt;/code&gt;. I ran it on two videos from a channel, and &lt;code&gt;done.txt&lt;/code&gt; was never created.&lt;/p&gt;

&lt;p&gt;yt-dlp only writes the archive when it downloads the video, and &lt;code&gt;--skip-download&lt;/code&gt; means it never does. You need one more flag:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;yt-dlp &lt;span class="nt"&gt;--skip-download&lt;/span&gt; &lt;span class="nt"&gt;--write-subs&lt;/span&gt; &lt;span class="nt"&gt;--write-auto-subs&lt;/span&gt; &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;--sub-langs&lt;/span&gt; en &lt;span class="nt"&gt;--sub-format&lt;/span&gt; srt &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;--download-archive&lt;/span&gt; &lt;span class="k"&gt;done&lt;/span&gt;.txt &lt;span class="nt"&gt;--force-write-archive&lt;/span&gt; &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;--sleep-subtitles&lt;/span&gt; 2 &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="nt"&gt;-o&lt;/span&gt; &lt;span class="s2"&gt;"subs/%(upload_date)s %(title).80s [%(id)s].%(ext)s"&lt;/span&gt; &lt;span class="se"&gt;\&lt;/span&gt;
  &lt;span class="s2"&gt;"https://www.youtube.com/@veritasium/videos"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The second run printed &lt;code&gt;has already been recorded in the archive&lt;/code&gt; for both videos and skipped them.&lt;/p&gt;

&lt;p&gt;The other flags in there:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;--sleep-subtitles 2&lt;/code&gt; waits between subtitle requests. On a big channel this is what keeps you clear of trap 1's 429.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;-I 1:50&lt;/code&gt; (not shown) limits the run to the newest 50 videos, which is handy for testing.&lt;/li&gt;
&lt;li&gt;The output template puts the date first, so files sort chronologically, and keeps the video id in brackets so you can always find the source.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  When yt-dlp stops being the right tool
&lt;/h2&gt;

&lt;p&gt;All of this works from a laptop on a home connection. From a cloud server it usually stops. YouTube blocks most datacenter IP ranges, and you start seeing &lt;code&gt;Sign in to confirm you're not a bot&lt;/code&gt;. Then you're dealing with cookies, residential proxies and PO tokens. I wrote about that in &lt;a href="https://dev.to/rapidshortsai_012681b0e02/youtube-transcript-api-works-locally-but-gets-blocked-on-your-server-why-and-3-fixes-181n"&gt;an earlier post&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;I build &lt;a href="https://bulktranscripts.co/youtube-subtitle-downloader" rel="noopener noreferrer"&gt;BulkTranscripts&lt;/a&gt;, which runs yt-dlp behind residential proxies so you don't have to. You paste a channel or playlist and get SRT, VTT, TXT or JSON for up to 1,000 videos. It's paid after the first 30, so if the commands above work for you, keep using them. They're free.&lt;/p&gt;

&lt;p&gt;If you hit a yt-dlp subtitle error that isn't covered here, paste it in the comments and I'll take a look.&lt;/p&gt;

</description>
      <category>python</category>
      <category>youtube</category>
      <category>cli</category>
      <category>tutorial</category>
    </item>
    <item>
      <title>youtube-transcript-api works locally but gets blocked on your server: why, and 3 fixes</title>
      <dc:creator>BulkTranscripts youtube </dc:creator>
      <pubDate>Tue, 27 Oct 2026 16:06:25 +0000</pubDate>
      <link>https://dev.to/rapidshortsai_012681b0e02/youtube-transcript-api-works-locally-but-gets-blocked-on-your-server-why-and-3-fixes-181n</link>
      <guid>https://dev.to/rapidshortsai_012681b0e02/youtube-transcript-api-works-locally-but-gets-blocked-on-your-server-why-and-3-fixes-181n</guid>
      <description>&lt;p&gt;You build something with &lt;a href="https://github.com/jdepoix/youtube-transcript-api" rel="noopener noreferrer"&gt;youtube-transcript-api&lt;/a&gt;. It works perfectly on your laptop. You deploy it to AWS, Render, Railway, Vercel or a VPS, and every call now fails with this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;youtube_transcript_api._errors.RequestBlocked: Could not retrieve a transcript for the video ...
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Sometimes it shows up as &lt;code&gt;IpBlocked&lt;/code&gt; instead. Your code didn't change and neither did the video, so what happened?&lt;/p&gt;

&lt;h2&gt;
  
  
  Why it happens
&lt;/h2&gt;

&lt;p&gt;YouTube blocks most IP addresses that belong to cloud providers. The library's own README says so: requests from AWS, Google Cloud, Azure and similar networks are very likely to be refused. A home IP can get blocked too if it sends a lot of requests in a short time.&lt;/p&gt;

&lt;p&gt;So the problem is &lt;strong&gt;where your request comes from&lt;/strong&gt;, not your code. That's also why switching libraries rarely helps. &lt;code&gt;yt-dlp&lt;/code&gt; on the same server usually hits the same wall, as a "Sign in to confirm you're not a bot" error.&lt;/p&gt;

&lt;p&gt;There are three ways around it.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fix 1: Rotating residential proxies
&lt;/h2&gt;

&lt;p&gt;This is the fix the library supports itself. Residential proxies send your request through ordinary home connections, which YouTube treats like normal viewers.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;youtube_transcript_api&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;YouTubeTranscriptApi&lt;/span&gt;
&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;youtube_transcript_api.proxies&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;WebshareProxyConfig&lt;/span&gt;

&lt;span class="n"&gt;ytt_api&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="nc"&gt;YouTubeTranscriptApi&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
    &lt;span class="n"&gt;proxy_config&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="nc"&gt;WebshareProxyConfig&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
        &lt;span class="n"&gt;proxy_username&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;&amp;lt;proxy-username&amp;gt;&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="n"&gt;proxy_password&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;&amp;lt;proxy-password&amp;gt;&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="p"&gt;)&lt;/span&gt;

&lt;span class="n"&gt;transcript&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;ytt_api&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;fetch&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;dQw4w9WgXcQ&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="nf"&gt;print&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt; &lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;join&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;s&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;text&lt;/span&gt; &lt;span class="k"&gt;for&lt;/span&gt; &lt;span class="n"&gt;s&lt;/span&gt; &lt;span class="ow"&gt;in&lt;/span&gt; &lt;span class="n"&gt;transcript&lt;/span&gt;&lt;span class="p"&gt;))&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Any provider works through &lt;code&gt;GenericProxyConfig&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;youtube_transcript_api.proxies&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;GenericProxyConfig&lt;/span&gt;

&lt;span class="n"&gt;ytt_api&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="nc"&gt;YouTubeTranscriptApi&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
    &lt;span class="n"&gt;proxy_config&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="nc"&gt;GenericProxyConfig&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
        &lt;span class="n"&gt;http_url&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;http://user:pass@proxy.example.com:8000&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="n"&gt;https_url&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;http://user:pass@proxy.example.com:8000&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="p"&gt;)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Things to know:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Buy &lt;strong&gt;rotating&lt;/strong&gt; residential proxies. Static or datacenter proxies get banned after a while.&lt;/li&gt;
&lt;li&gt;You pay per GB, and the price varies a lot between providers. Transcripts are small, but YouTube pages around them are not.&lt;/li&gt;
&lt;li&gt;Some requests will still fail, so add retries with backoff.&lt;/li&gt;
&lt;li&gt;Cache every transcript you fetch. Captions rarely change, and a cache hit costs no proxy traffic.
&lt;/li&gt;
&lt;/ul&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;json&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;pathlib&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;time&lt;/span&gt;
&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;youtube_transcript_api&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;YouTubeTranscriptApi&lt;/span&gt;
&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;youtube_transcript_api._errors&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;RequestBlocked&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;IpBlocked&lt;/span&gt;

&lt;span class="n"&gt;CACHE&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;pathlib&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nc"&gt;Path&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;transcripts&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="n"&gt;CACHE&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;mkdir&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;exist_ok&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="bp"&gt;True&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;

&lt;span class="k"&gt;def&lt;/span&gt; &lt;span class="nf"&gt;get_transcript&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;api&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;YouTubeTranscriptApi&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;video_id&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="nb"&gt;str&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;tries&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="nb"&gt;int&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="mi"&gt;4&lt;/span&gt;&lt;span class="p"&gt;):&lt;/span&gt;
    &lt;span class="n"&gt;path&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;CACHE&lt;/span&gt; &lt;span class="o"&gt;/&lt;/span&gt; &lt;span class="sa"&gt;f&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="si"&gt;{&lt;/span&gt;&lt;span class="n"&gt;video_id&lt;/span&gt;&lt;span class="si"&gt;}&lt;/span&gt;&lt;span class="s"&gt;.json&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;
    &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="n"&gt;path&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;exists&lt;/span&gt;&lt;span class="p"&gt;():&lt;/span&gt;
        &lt;span class="k"&gt;return&lt;/span&gt; &lt;span class="n"&gt;json&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;loads&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;path&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;read_text&lt;/span&gt;&lt;span class="p"&gt;())&lt;/span&gt;
    &lt;span class="k"&gt;for&lt;/span&gt; &lt;span class="n"&gt;attempt&lt;/span&gt; &lt;span class="ow"&gt;in&lt;/span&gt; &lt;span class="nf"&gt;range&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;tries&lt;/span&gt;&lt;span class="p"&gt;):&lt;/span&gt;
        &lt;span class="k"&gt;try&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
            &lt;span class="n"&gt;snippets&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
                &lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;text&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;s&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;text&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;start&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;s&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;start&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;duration&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;s&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;duration&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;
                &lt;span class="k"&gt;for&lt;/span&gt; &lt;span class="n"&gt;s&lt;/span&gt; &lt;span class="ow"&gt;in&lt;/span&gt; &lt;span class="n"&gt;api&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;fetch&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;video_id&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
            &lt;span class="p"&gt;]&lt;/span&gt;
            &lt;span class="n"&gt;path&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;write_text&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;json&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;dumps&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;snippets&lt;/span&gt;&lt;span class="p"&gt;))&lt;/span&gt;
            &lt;span class="k"&gt;return&lt;/span&gt; &lt;span class="n"&gt;snippets&lt;/span&gt;
        &lt;span class="nf"&gt;except &lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;RequestBlocked&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;IpBlocked&lt;/span&gt;&lt;span class="p"&gt;):&lt;/span&gt;
            &lt;span class="n"&gt;time&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;sleep&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="mi"&gt;2&lt;/span&gt; &lt;span class="o"&gt;**&lt;/span&gt; &lt;span class="n"&gt;attempt&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;  &lt;span class="c1"&gt;# rotating proxies give a new IP on retry
&lt;/span&gt;    &lt;span class="k"&gt;raise&lt;/span&gt; &lt;span class="nc"&gt;RuntimeError&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sa"&gt;f&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;still blocked after &lt;/span&gt;&lt;span class="si"&gt;{&lt;/span&gt;&lt;span class="n"&gt;tries&lt;/span&gt;&lt;span class="si"&gt;}&lt;/span&gt;&lt;span class="s"&gt; tries: &lt;/span&gt;&lt;span class="si"&gt;{&lt;/span&gt;&lt;span class="n"&gt;video_id&lt;/span&gt;&lt;span class="si"&gt;}&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;Good fit:&lt;/strong&gt; you want full control, you fetch a lot every day, and you're fine maintaining it.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fix 2: Fetch from a machine that isn't in a data center
&lt;/h2&gt;

&lt;p&gt;If your volume is small, run the fetching part somewhere with a normal home IP: an old laptop, a Raspberry Pi, or your dev machine. Your cloud app puts video IDs in a queue (a database table is enough), and the home worker fetches the transcripts and writes them back.&lt;/p&gt;

&lt;p&gt;It's free, but it's another machine to keep running. If you send too many requests, your home IP can get blocked too. Keep it slow, a few requests a minute, and cache everything.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Good fit:&lt;/strong&gt; side projects, a few dozen videos a day.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fix 3: Use a hosted transcript API
&lt;/h2&gt;

&lt;p&gt;The third option is to make the IP problem somebody else's. Several hosted APIs do this: Supadata, a few Apify actors, and the one I build, &lt;a href="https://bulktranscripts.co/youtube-transcript-api" rel="noopener noreferrer"&gt;BulkTranscripts&lt;/a&gt;. They run the proxies and retries, and your server makes a plain HTTPS call that YouTube never sees.&lt;/p&gt;

&lt;p&gt;Disclosure: I build BulkTranscripts, so here's what it looks like:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;requests&lt;/span&gt;

&lt;span class="n"&gt;r&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;requests&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;get&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
    &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;https://bulktranscripts.co/api/v1/transcript&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="n"&gt;params&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;video&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;https://www.youtube.com/watch?v=dQw4w9WgXcQ&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;},&lt;/span&gt;
    &lt;span class="n"&gt;headers&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;Authorization&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;Bearer bt_ak_YOUR_KEY&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;},&lt;/span&gt;
    &lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;60&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="n"&gt;r&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;raise_for_status&lt;/span&gt;&lt;span class="p"&gt;()&lt;/span&gt;
&lt;span class="n"&gt;data&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;r&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;json&lt;/span&gt;&lt;span class="p"&gt;()&lt;/span&gt;
&lt;span class="nf"&gt;print&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;data&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;title&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;],&lt;/span&gt; &lt;span class="n"&gt;data&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;word_count&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;])&lt;/span&gt;
&lt;span class="nf"&gt;print&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;data&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;text&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;][:&lt;/span&gt;&lt;span class="mi"&gt;500&lt;/span&gt;&lt;span class="p"&gt;])&lt;/span&gt;          &lt;span class="c1"&gt;# full transcript as one string
&lt;/span&gt;&lt;span class="nf"&gt;print&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;data&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;segments&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;][&lt;/span&gt;&lt;span class="mi"&gt;0&lt;/span&gt;&lt;span class="p"&gt;])&lt;/span&gt;         &lt;span class="c1"&gt;# {"text": ..., "start": ..., "duration": ...}
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Some details, since you'd want them before choosing:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;30 free credits to try it, no card. After that, packs start at $4.99 for 200 transcripts, and credits don't expire.&lt;/li&gt;
&lt;li&gt;A failed fetch isn't charged, and re-reading a transcript you already fetched is free.&lt;/li&gt;
&lt;li&gt;There are endpoints for a whole channel or playlist (&lt;code&gt;/api/v1/channel/videos&lt;/code&gt;, &lt;code&gt;/api/v1/playlist/videos&lt;/code&gt;), plus an MCP server if you want Claude or ChatGPT to fetch transcripts directly.&lt;/li&gt;
&lt;li&gt;
&lt;a href="https://bulktranscripts.co/docs" rel="noopener noreferrer"&gt;Docs here&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Whichever hosted API you pick, check three things: what a failed request costs, whether re-fetches are billed, and whether it can do channels and playlists in bulk.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Good fit:&lt;/strong&gt; you want to ship your actual product and stop thinking about proxies.&lt;/p&gt;

&lt;h2&gt;
  
  
  Which one to pick
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Your situation&lt;/th&gt;
&lt;th&gt;Pick&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Hobby project, under ~50 videos a day&lt;/td&gt;
&lt;td&gt;Fix 2: a home worker&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;High volume, you want control and can maintain it&lt;/td&gt;
&lt;td&gt;Fix 1: rotating residential proxies&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;You want it working today and don't want to maintain it&lt;/td&gt;
&lt;td&gt;Fix 3: a hosted API&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Whatever you choose, &lt;strong&gt;cache transcripts&lt;/strong&gt;. It's the one thing that makes all three cheaper and more reliable.&lt;/p&gt;

&lt;p&gt;If you've found another way around &lt;code&gt;RequestBlocked&lt;/code&gt;, I'd like to hear it in the comments.&lt;/p&gt;

</description>
      <category>python</category>
      <category>youtube</category>
      <category>api</category>
      <category>webdev</category>
    </item>
    <item>
      <title>I gave ChatGPT YouTube transcripts through MCP. Here is how it works</title>
      <dc:creator>BulkTranscripts youtube </dc:creator>
      <pubDate>Tue, 22 Sep 2026 17:00:38 +0000</pubDate>
      <link>https://dev.to/rapidshortsai_012681b0e02/i-gave-chatgpt-youtube-transcripts-through-mcp-here-is-how-it-works-27co</link>
      <guid>https://dev.to/rapidshortsai_012681b0e02/i-gave-chatgpt-youtube-transcripts-through-mcp-here-is-how-it-works-27co</guid>
      <description>&lt;p&gt;Every few weeks I need to read a lot of YouTube. Not watch, read. A conference channel drops twenty talks, a framework maintainer explains a breaking change on a podcast, a competitor ships something and walks through it on camera. The information is in the audio, and the audio is slow.&lt;/p&gt;

&lt;p&gt;For a long time my workaround was the "show transcript" panel on YouTube, copy, paste into ChatGPT, repeat. It works for one video. It falls apart at five. And the pasted text is full of timestamps and line breaks that eat context.&lt;/p&gt;

&lt;p&gt;So I built the thing I wanted: an MCP server that hands an AI model clean transcripts, channel search, and playlist listings as tools. It's called &lt;a href="https://bulktranscripts.co" rel="noopener noreferrer"&gt;BulkTranscripts&lt;/a&gt; and this post is how it works under the hood, plus how to wire it into whatever client you use.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the model gets
&lt;/h2&gt;

&lt;p&gt;The server exposes a small set of tools. The ones I use most:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;get_transcript&lt;/code&gt; for one video, returned as plain text with no timestamps unless you ask for them&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;get_transcripts&lt;/code&gt; for a batch of video IDs in one call&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;search_channel&lt;/code&gt; to find videos on a channel by topic&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;get_latest_videos&lt;/code&gt; to see what a channel uploaded recently, always free&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;start_bulk_extract&lt;/code&gt; for a whole channel or playlist, up to a thousand videos, running server-side while the model polls &lt;code&gt;get_bulk_status&lt;/code&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That last one matters more than it sounds. A model that has to fetch 200 transcripts inline will time out or blow its context. A detached job that fills a library on the server, with the model paging through results afterwards, is the difference between a demo and something you can use on a real channel.&lt;/p&gt;

&lt;h2&gt;
  
  
  Wiring it up
&lt;/h2&gt;

&lt;p&gt;If your client speaks OAuth (ChatGPT, Claude web and desktop, Claude Code, Cursor, VS Code), you add the endpoint and sign in with Google. No key to manage.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;https://bulktranscripts.co/mcp
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;In Claude Code that's one line:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;claude mcp add &lt;span class="nt"&gt;--transport&lt;/span&gt; http bulktranscripts https://bulktranscripts.co/mcp
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;In ChatGPT it lives under Apps, and the sign-in is a normal Google prompt. Thirty transcripts come free with the account, no card.&lt;/p&gt;

&lt;p&gt;For clients without OAuth, or for scripts, you create an API key in the dashboard and send it as a bearer token:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"mcpServers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="nl"&gt;"bulktranscripts"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"url"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"https://bulktranscripts.co/mcp"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"headers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"Authorization"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"Bearer bt_ak_..."&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The setup page for each client is in the &lt;a href="https://bulktranscripts.co/docs#mcp" rel="noopener noreferrer"&gt;docs&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  The credit rule, because it trips people up
&lt;/h2&gt;

&lt;p&gt;A transcript costs one credit the first time it enters your library. After that, reading it again is free, forever, from any client. Live lookups like channel search cost one credit. &lt;code&gt;get_latest_videos&lt;/code&gt; is free so you can poll it on a schedule and only fetch what's new.&lt;/p&gt;

&lt;p&gt;Credits follow the account, not the connector. Buying a pack never touches your MCP config. If a batch runs past your balance, you get the transcripts that completed and a clear count of how many were stopped, so the model can tell you instead of silently truncating.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I actually do with it
&lt;/h2&gt;

&lt;p&gt;The workflow that made me build this: pick one channel, ask for the fifteen most relevant videos on a topic, pull all the transcripts, and have the model produce something structured from them. A changelog digest. A "what did they claim and where" table with video attribution. A playbook from a series of talks.&lt;/p&gt;

&lt;p&gt;Here's the prompt shape, in ChatGPT with the connector on:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Look through @channel's recent videos about [topic]. Use the 15 most relevant. Pull the transcripts and build me a practical guide, citing which video each idea comes from.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;The model makes a handful of tool calls, thinks for a couple of minutes, and the output is grounded in what was said rather than in what the titles suggest. That grounding is the whole point. Titles are marketing. Transcripts are the talk.&lt;/p&gt;

&lt;h2&gt;
  
  
  Things I'd warn you about
&lt;/h2&gt;

&lt;p&gt;Transcripts are only as good as YouTube's captions. Auto-generated captions on a heavy accent or a noisy room come through rough. Uploader-provided captions are great. You can't tell which you got until you read it.&lt;/p&gt;

&lt;p&gt;Also, don't let the model summarize. Ask it for claims, decisions, and quotes with attribution. Summaries flatten the interesting parts. Claims give you something to check or disagree with.&lt;/p&gt;

&lt;p&gt;If you try it and something breaks, tell me. The endpoint is &lt;a href="https://bulktranscripts.co" rel="noopener noreferrer"&gt;bulktranscripts.co/mcp&lt;/a&gt;, and the free credits are enough to run the whole workflow above once or twice before you decide if it's worth keeping.&lt;/p&gt;

</description>
      <category>chatgpt</category>
      <category>mcp</category>
      <category>ai</category>
      <category>youtube</category>
    </item>
  </channel>
</rss>
