[YouTube] Intermittent "HTTP Error 403: Forbidden" when downloading videos

[BillyCool]

Checklist

• I've read the NO AI / NO LLM POLICY and understand I may be blocked from the repository if it is violated
• I'm reporting that yt-dlp is broken on a supported site
• I've verified that I have updated yt-dlp to nightly or master (update instructions)
• I've checked that all provided URLs are playable in a browser with the same IP and same login details
• I've checked that all URLs and arguments with special characters are properly quoted or escaped
• I've searched known issues, the FAQ, and the bugtracker for similar issues including closed ones. DO NOT post duplicates
• I've read about sharing account credentials and I'm willing to share it if required

Region

Australia

Provide a description that is worded well enough to be understood

Lately noticed a lot of my videos intermittently fail to download with ERROR: unable to download video data: HTTP Error 403: Forbidden. The same videos will download successfully after a few retries without me changing anything. Issue is present on both Windows and Linux (Docker), with VPN on or off (off here), account cookies on or off (off here) has been going on for a few weeks now.

Tried searching through similar issues raised but didn't find any working solutions. I am not using -vU because my app that integrates with yt-dlp checks for updates on every startup so checking again on every download is redundant. I am on the latest nightly version at the time of writing. Yes, the video is playable on my browser, with and without account, across multiple browsers.

I am attached some log files for recent single video, each log is a download attempt via a separate terminal/process, attempts happened back to back. Only thing altered in the logs is my IP for safety. One thing that stood out to me is that all failing logs have this query parameter siu=1 but I don't know what it is, maybe it helps someone who does.

Example of a video with account cookies off:
iQZ81MbjKpU-20260808-195246-failed-redacted.log
iQZ81MbjKpU-20260808-195607-success-redacted.log

Example of a different video with account cookies on:
rg3aBTjsUms-20260808-200652-failed-redacted.log
rg3aBTjsUms-20260808-200734-failed-redacted.log
rg3aBTjsUms-20260808-200910-success-redacted.log

Can provide more logs if needed. Thanks in advance!

Provide verbose output that clearly demonstrates the problem

• Run your yt-dlp command with -vU flag added (yt-dlp -vU <your command line>)
• If using API, add 'verbose': True to YoutubeDL params instead
• Copy the WHOLE output (starting with [debug] Command-line config) and insert it below

Complete Verbose Output

[debug] Command-line config: ['--extractor-args', 'youtubepot-bgutilhttp:base_url=http://host.docker.internal:4416', '-P', '/media/youtube/People/LaurieWired', '-P', 'temp:/media/youtube/.temp', '-S', 'res:1080', '-f', 'bv*[fps<=60]+ba/b', 'https://www.youtube.com/watch?v=iQZ81MbjKpU']
[debug] System config "/etc/yt-dlp.conf": ['--no-js-runtimes', '--js-runtimes', 'node', '--live-from-start', '--no-playlist', '--concurrent-fragments', '4', '--output', '%(channel)s - %(upload_date>%Y-%m-%d)s - %(title)s.%(ext)s', '--output', 'subtitle:Subtitles/%(id)s.%(ext)s', '--output', 'thumbnail:Thumbnails/%(id)s.%(ext)s', '--windows-filenames', '--force-overwrites', '--write-thumbnail', '--write-subs', '--write-auto-subs', '--sub-format', 'vtt', '--sub-langs', '.*-orig', '--embed-subs', '--embed-thumbnail', '--embed-metadata', '--embed-chapters', '--extractor-args', 'youtube:player-client=mweb', '-t', 'sleep', '--force-ipv4', '--verbose']
[debug] Encodings: locale utf-8, fs utf-8, pref utf-8, out utf-8 (No ANSI), error utf-8 (No ANSI), screen utf-8 (No ANSI)
[debug] yt-dlp version nightly@2026.08.04.234419 from yt-dlp/yt-dlp-nightly-builds (linux_exe)
[debug] Python 3.14.6 (CPython x86_64 64bit) - Linux-6.6.87.2-microsoft-standard-WSL2-x86_64-with-glibc2.39 (OpenSSL 3.5.7 9 Jun 2026, glibc 2.39)
[debug] exe versions: ffmpeg 6.1.1 (setts), ffprobe 6.1.1
[debug] Optional libraries: Cryptodome-3.23.0, brotli-1.2.0, certifi-2026.06.17, curl_cffi-0.15.0, mutagen-1.48.1, requests-2.34.2, secretstorage-3.5.0, sqlite3-3.53.3, urllib3-2.7.0, websockets-16.0, yt_dlp_ejs-0.8.0
[debug] JS runtimes: node-22.23.2
[debug] Proxy map: {}
[debug] Request Handlers: urllib, requests, websockets, curl_cffi
[debug] Plugin directories: /etc/yt-dlp/plugins/bgutil-ytdlp-pot-provider/yt_dlp_plugins
[debug] Loaded 1744 extractors
[debug] [youtube] [pot:bgutil:script-node] No server_home or script_path passed, defaulting to /root/bgutil-ytdlp-pot-provider/server
[debug] [youtube] [pot:bgutil:script-node] Script path doesn't exist: /root/bgutil-ytdlp-pot-provider/server/build/generate_once.js
[debug] [youtube] [pot:bgutil:script-deno] No server_home or script_path passed, defaulting to /root/bgutil-ytdlp-pot-provider/server
[debug] [youtube] [pot:bgutil:script-deno] Script path doesn't exist: /root/bgutil-ytdlp-pot-provider/server/src/generate_once.ts
[debug] [youtube] [pot] PO Token Providers: bgutil:http-1.3.1 (external), bgutil:script-node-1.3.1 (external, unavailable), bgutil:script-deno-1.3.1 (external, unavailable)
[debug] [youtube] [pot] PO Token Cache Providers: memory
[debug] [youtube] [pot] PO Token Cache Spec Providers: webpo
[debug] [youtube] [jsc] JS Challenge Providers: bun (unavailable), deno (unavailable), node, quickjs (unavailable)
[youtube] Extracting URL: https://www.youtube.com/watch?v=iQZ81MbjKpU
[youtube] iQZ81MbjKpU: Downloading webpage
[youtube] Sleeping 0.75 seconds ...
[youtube] iQZ81MbjKpU: Downloading mweb client config
[debug] [youtube] Forcing "main" player JS variant for player 854a788e
        original url = /s/player/854a788e/player_es6.vflset/en_US/base.js
[debug] [youtube] iQZ81MbjKpU: Detected experiment to bind GVS PO Token to video ID for mweb client
[youtube] Sleeping 0.75 seconds ...
[youtube] iQZ81MbjKpU: Downloading mweb player API JSON
[debug] [youtube] iQZ81MbjKpU: Detected a 6s ad skippable after 5s for mweb
[youtube] [pot:bgutil:http] Generating a gvs PO Token for mweb client via bgutil HTTP server
[debug] [youtube] iQZ81MbjKpU: Retrieved a gvs PO Token for mweb client
[youtube] Sleeping 0.75 seconds ...
[youtube] iQZ81MbjKpU: Downloading player 854a788e-main
[youtube] [jsc:node] Solving JS challenges using node
[debug] [youtube] [jsc:node] Using challenge solver lib script v0.8.0 (source: python package, variant: minified)
[debug] [youtube] [jsc:node] Using challenge solver core script v0.8.0 (source: python package, variant: minified)
[debug] [youtube] [jsc:node] Running node: node --experimental-permission --no-warnings=ExperimentalWarning -
[debug] [youtube] iQZ81MbjKpU: Retrieved a subs PO Token for mweb client
[info] iQZ81MbjKpU: Downloading subtitles: en-orig
[debug] Sort order given by user: res:1080
[debug] Sort order given by extractor: quality, res, fps, hdr:12, source, vcodec, channels, acodec, lang, proto
[debug] Formats sorted by: hasvid, ie_pref, res:1080(1080.0), quality, fps, hdr:12(7), source, vcodec, channels, acodec, lang, proto, size, br, asr, vext, aext, hasaud, id
[info] iQZ81MbjKpU: Downloading 1 format(s): 399+251
[info] Writing video subtitles to: /media/youtube/.temp/LaurieWired - 2025-07-01 - grep isn't what you think it means....en-orig.vtt
[debug] Invoking http downloader on "https://www.youtube.com/api/timedtext?v=iQZ81MbjKpU&ei=8Ih3aoScH_TfssUPva_H4Qc&caps=asr&opi=112496729&exp=xpe&xoaf=5&xowf=1&hl=en&ip=0.0.0.0&ipbits=0&expire=1786243936&sparams=ip%2Cipbits%2Cexpire%2Cv%2Cei%2Ccaps%2Copi%2Cexp%2Cxoaf&signature=A7915111CB91444BC076EB29AC14CB3271BEA94B.D062EE4886EB404406DB20840772F3CE1560C430&key=yt8&kind=asr&lang=en&pot=MlsA8FSRn4MzegC0VOtZTPDY48tUPJsr87OVfGiOGEdJCj3HhE-B90BU6Z2q9q4AukKbXzDPU0zFmlBLMWO5U86lQppTAyioZ9xdjqNi8XJQKt6sTYd6ggqp55Di&potc=1&c=MWEB&fmt=vtt"
[download] Sleeping 5.00 seconds ...
[download] Destination: /media/youtube/.temp/LaurieWired - 2025-07-01 - grep isn't what you think it means....en-orig.vtt

[download]    1.00KiB at  299.04KiB/s (00:00:00)
[download]    3.00KiB at  748.85KiB/s (00:00:00)
[download]    7.00KiB at    1.55MiB/s (00:00:00)
[download]   15.00KiB at    2.94MiB/s (00:00:00)
[download]   31.00KiB at    5.50MiB/s (00:00:00)
[download]   63.00KiB at    9.53MiB/s (00:00:00)
[download]   98.29KiB at   10.25MiB/s (00:00:00)
[download] 100% of   98.29KiB in 00:00:00 at 349.43KiB/s
[info] Downloading video thumbnail 42 ...
[info] Writing video thumbnail 42 to: /media/youtube/.temp/LaurieWired - 2025-07-01 - grep isn't what you think it means....webp
WARNING: webm doesn't support embedding a thumbnail, mkv will be used
[download] Sleeping 16.52 seconds ...
[debug] Invoking http downloader on "https://rr1---sn-ajhxac-5c.googlevideo.com/videoplayback?expire=1786240336&ei=8Ih3aoScH_TfssUPva_H4Qc&ip=192.168.1.1&id=o-ADvkTzlR8CQWJJbj1ZefmoPyE86t8EpbtOaDc4AwOiTz&itag=399&aitags=133%2C134%2C135%2C136%2C137%2C160%2C242%2C243%2C244%2C247%2C248%2C271%2C278%2C313%2C394%2C395%2C396%2C397%2C398%2C399%2C400%2C401%2C598&source=youtube&requiressl=yes&xpc=EgVo2aDSNQ%3D%3D&cps=246&met=1786218736%2C&mh=Y3&mm=31%2C26&mn=sn-ajhxac-5c%2Csn-u1ntq5-52&ms=au%2Conr&mv=m&mvi=1&pl=22&rms=au%2Cau&initcwndbps=2720000&bui=AZFlqhOQ4RUQqr3EzxbnUY4dOtMPmolGP1kpvbgV7b10i8FN7iiClhOh0wiJB0lUX0mkhdn6GGx7kfIT&spc=KBGBcq-1KpjvlFcgXU0GZY-VK0DBuSuvNnE6f5v-C8cJFKjpiXlfC4-7_aFahxjc274d1qBoDq65lkJEOfzxDlJnx2c&vprv=1&svpuc=1&mime=video%2Fmp4&ns=8qPTRflZg6SFsJ7qMNxNCO8X&rqh=1&gir=yes&clen=82990296&dur=692.358&lmt=1751625787822220&mt=1786218300&fvip=1&keepalive=yes&fexp=51565116%2C51973818%2C52089683&c=MWEB&sefc=1&txp=5537534&n=SxGPabqXCFXbGw&sparams=expire%2Cei%2Cip%2Cid%2Caitags%2Csource%2Crequiressl%2Cxpc%2Cbui%2Cspc%2Cvprv%2Csvpuc%2Cmime%2Cns%2Crqh%2Cgir%2Cclen%2Cdur%2Clmt&sig=AE0s2JYwRAIgXTA6FJ-cI_h0seVPdDJ2Uyrmlq4ZXUssNnwFIYRTLfACICYM5dME3b9hsH-WdxF6AOoeRNx9mkKAQ4RrEwqkeNtR&lsparams=cps%2Cmet%2Cmh%2Cmm%2Cmn%2Cms%2Cmv%2Cmvi%2Cpl%2Crms%2Cinitcwndbps&lsig=APaTxxMwRgIhAIcaF3vBJY7GeOftB3WUorKX3iCle_cnKYwxdAZqYFKVAiEAqhrsvtwRfEfS5xvrai_qaIg3IeqYD5rgwn57reAfB-Y%3D&pot=MlsA8FSRn4MzegC0VOtZTPDY48tUPJsr87OVfGiOGEdJCj3HhE-B90BU6Z2q9q4AukKbXzDPU0zFmlBLMWO5U86lQppTAyioZ9xdjqNi8XJQKt6sTYd6ggqp55Di"
ERROR: unable to download video data: HTTP Error 403: Forbidden
Traceback (most recent call last):
  File "yt_dlp/YoutubeDL.py", line 3571, in process_info
  File "yt_dlp/YoutubeDL.py", line 3324, in dl
  File "yt_dlp/downloader/common.py", line 480, in download
  File "yt_dlp/downloader/http.py", line 362, in real_download
  File "yt_dlp/downloader/http.py", line 119, in establish_connection
  File "yt_dlp/YoutubeDL.py", line 4305, in urlopen
  File "yt_dlp/networking/common.py", line 117, in send
  File "yt_dlp/networking/_helper.py", line 194, in wrapper
  File "yt_dlp/networking/common.py", line 359, in send
  File "yt_dlp/networking/_requests.py", line 361, in _send
yt_dlp.networking.exceptions.HTTPError: HTTP Error 403: Forbidden

[coletdjnz]

Might be #17368, which is an issue with the bgutil pot provider.

Does it always fail with a 403 immediately or does it sometimes start downloading then fail?

(also maybe related: #17389)

[BillyCool]

Does it always fail with a 403 immediately or does it sometimes start downloading then fail?

It feels like 50/50, I'll typically queue a few videos at the same time, half will go through and download fine first try, half will error out and need to be retried a few times.

[doc-txt]

It feels like 50/50

This is #17368
Brainicism/bgutil-ytdlp-pot-provider#243
They’ll probably fix the POT soon. I tried downloading it a few times—no 403 errors so far.

[MalwareCreater]

I was facing this for about a week.

It seems to be an IP locking on videos from YT's side.

Solution: Keep the video open/keep it playing in a browser tab and then try to download it. It will work without that error. It has worked in 100% of my videos, both public and restricted ones.

[BillyCool]

Solution: Keep the video open/keep it playing in a browser tab and then try to download it. It will work without that error. It has worked in 100% of my videos, both public and restricted ones.

This hasn't worked for me unfortunately. I've tried rotating my home IP and using non-blocked VPN IPs and it's a similar result every time.

This is #17368

I think this is correct. It seems to be something with POT and the mweb client (possibly any client that works with POT). In case others are reading, what solved this for now was switching my client config to visionos,web_embedded,mweb instead of going straight to mweb (even though some docs recommend this setup). These clients don't need POT and download normally, for now 😄

[ruben1PvP]

In case others are reading, what solved this for now was switching my client config to visionos,web_embedded,mweb instead of going straight to mweb (even though some docs recommend this setup). These clients don't need POT and download normally, for now

visionos is a default client, and no default client needs po tokens. I don't know why would you need web_embedded, this client was only necessary when you needed more than the original language when visionos didn't exist, and to solve problems like #17389. You don't need any --extractor-args.

[illuminatibit]

Might be #17368, which is an issue with the bgutil pot provider.

Does it always fail with a 403 immediately or does it sometimes start downloading then fail?

(also maybe related: #17389)

Attention!
The problem isn't with tokens or cookies.
The android_vr player also gets a 403 error when downloading videos after receiving the metadata - the same video may not download.
And I don't use tokens or cookies; I use either a regular IP or a proxy, many proxies. It makes no difference, sometimes it works, sometimes it doesn't; you can check any video yourself.
I think the problem is much more serious and will soon be like this everywhere, and this is some kind of YouTube test... Or is it? Or is there something wrong with our code?
What do you think?

I'm panicking because I have a video download site and I need the 18 format, which isn't available in VisionOS (which currently works fine). TV_downgraded works fine without tokens, but it takes a long time to load, and my server can't handle it, there are so many people... My only normal option was android_vr, which now periodically gives people a 403, which is, of course, terrible...

[ruben1PvP]

Attention!
The problem isn't with tokens or cookies.
The android_vr player also gets a 403 error when downloading videos after receiving the metadata - the same video may not download.
And I don't use tokens or cookies; I use either a regular IP or a proxy, many proxies. It makes no difference, sometimes it works, sometimes it doesn't; you can check any video yourself.
I think the problem is much more serious and will soon be like this everywhere, and this is some kind of YouTube test... Or is it? Or is there something wrong with our code?
What do you think?

You are experimenting something similar to #16796 (the android_vr client getting experiments that make it mandatory to use po tokens). So yes, it is "some kind of Youtube test". That's why visionos exists now.

I'm panicking because I have a video download site and I need the 18 format, which isn't available in VisionOS (which currently works fine). TV_downgraded works fine without tokens, but it takes a long time to load, and my server can't handle it, there are so many people... My only normal option was android_vr, which now periodically gives people a 403, which is, of course, terrible...

If you only want the 18 format (and only the 18 format) and using deno or other javascript runtimes is too much for you, you could get it from the android client (--extractor-args "youtube:player-client=android")

[illuminatibit]

Attention!
The problem isn't with tokens or cookies.
The android_vr player also gets a 403 error when downloading videos after receiving the metadata - the same video may not download.
And I don't use tokens or cookies; I use either a regular IP or a proxy, many proxies. It makes no difference, sometimes it works, sometimes it doesn't; you can check any video yourself.
I think the problem is much more serious and will soon be like this everywhere, and this is some kind of YouTube test... Or is it? Or is there something wrong with our code?
What do you think?

You are experimenting something similar to #16796 (the android_vr client getting experiments that make it mandatory to use po tokens). So yes, it is "some kind of Youtube test". That's why visionos exists now.

I'm panicking because I have a video download site and I need the 18 format, which isn't available in VisionOS (which currently works fine). TV_downgraded works fine without tokens, but it takes a long time to load, and my server can't handle it, there are so many people... My only normal option was android_vr, which now periodically gives people a 403, which is, of course, terrible...

If you only want the 18 format (and only the 18 format) and using deno or other javascript runtimes is too much for you, you could get it from the android client (--extractor-args "youtube:player-client=android")

I didn't know that android_vr started giving 403 errors because of tokens (my site has had this issue for the last two days). I can implement them, but it seems easier to switch to the regular Android client now. It's strange. I tried using it before, but I didn't like it for some reason. Maybe it took a long time to find the metadata, but now it's faster.

Anyway, thanks for the support. I would never have considered using a regular Android client on my own.

I'm already using the node for the Java script.
But for some reason, TV_downgraded takes a long time to load for each user. Maybe my script doesn't reset something after using this player, and I need to reset it with the node?

I've been looking for other players for 18 format, but it's been no use. I've even considered installing PO tokens, but from this thread, I see that even PO tokens have problems...

Thanks!Respect!