TikTok videos can't be downloaded

[sondhg]

A few days ago I could still download TikTok videos with gallery-dl. Today I can't download any. Here's what I get after running gallery-dl --verbose "https://www.tiktok.com/@hakimdailyyy/video/7588853215685381384". I did provide the cookies file in my ~/.config/gallery-dl/config.json. What went wrong? Thank you.

[gallery-dl][debug] Version 1.31.1
[gallery-dl][debug] Python 3.13.11 - Linux-6.18.2-2-cachyos-x86_64-with-glibc2.42
[gallery-dl][debug] requests 2.32.5 - urllib3 2.6.2
[gallery-dl][debug] Configuration Files ['${HOME}/.config/gallery-dl/config.json']
[gallery-dl][debug] Starting DownloadJob for 'https://www.tiktok.com/@hakimdailyyy/video/7588853215685381384'
[tiktok][debug] Using TiktokPostExtractor for 'https://www.tiktok.com/@hakimdailyyy/video/7588853215685381384'
[tiktok][debug] cookies: Loading cookies from '$HOME/.config/gallery-dl/www.tiktok.com_cookies.txt'
[urllib3.connectionpool][debug] Starting new HTTPS connection (1): www.tiktok.com:443
[urllib3.connectionpool][debug] https://www.tiktok.com:443 "GET /@hakimdailyyy/video/7588853215685381384 HTTP/1.1" 200 640
[tiktok][warning] 7588853215685381384: Failed to retrieve rehydration data (1/4)
[tiktok][debug] Sleeping 30.00 seconds (retry)
[urllib3.connectionpool][debug] https://www.tiktok.com:443 "GET /@hakimdailyyy/video/7588853215685381384 HTTP/1.1" 200 643
[tiktok][warning] 7588853215685381384: Failed to retrieve rehydration data (2/4)
[tiktok][debug] Sleeping 30.00 seconds (retry)
[urllib3.connectionpool][debug] https://www.tiktok.com:443 "GET /@hakimdailyyy/video/7588853215685381384 HTTP/1.1" 200 642
[tiktok][warning] 7588853215685381384: Failed to retrieve rehydration data (3/4)
[tiktok][debug] Sleeping 30.00 seconds (retry)
[urllib3.connectionpool][debug] https://www.tiktok.com:443 "GET /@hakimdailyyy/video/7588853215685381384 HTTP/1.1" 200 642
[tiktok][warning] 7588853215685381384: Failed to retrieve rehydration data (4/4)
[tiktok][debug] Sleeping 30.00 seconds (retry)
[urllib3.connectionpool][debug] https://www.tiktok.com:443 "GET /@hakimdailyyy/video/7588853215685381384 HTTP/1.1" 200 643
[tiktok][error] An unexpected error occurred: JSONDecodeError - Expecting value: line 1 column 1 (char 0). Please run gallery-dl again with the --verbose flag, copy its output and report this issue on https://github.com/mikf/gallery-dl/issues .
[tiktok][debug] 
Traceback (most recent call last):
  File "/home/sondhg/.local/share/uv/tools/gallery-dl/lib/python3.13/site-packages/gallery_dl/job.py", line 158, in run
    msg = self.dispatch(extractor)
  File "/home/sondhg/.local/share/uv/tools/gallery-dl/lib/python3.13/site-packages/gallery_dl/job.py", line 209, in dispatch
    for msg, url, kwdict in messages:
                            ^^^^^^^^
  File "/home/sondhg/.local/share/uv/tools/gallery-dl/lib/python3.13/site-packages/gallery_dl/extractor/tiktok.py", line 33, in items
    data = self._extract_rehydration_data(tiktok_url)
  File "/home/sondhg/.local/share/uv/tools/gallery-dl/lib/python3.13/site-packages/gallery_dl/extractor/tiktok.py", line 118, in _extract_rehydration_data
    return util.json_loads(data)["__DEFAULT_SCOPE__"]
           ~~~~~~~~~~~~~~~^^^^^^
  File "/usr/lib/python3.13/json/decoder.py", line 345, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
               ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.13/json/decoder.py", line 363, in raw_decode
    raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Edit: I installed gallery-dl via uv tool install gallery-dl --with 'yt-dlp[default,curl-cffi]'. It has been working fine for months.

Just now I tried

uv tool install yt-dlp --with curl-cffi
yt-dlp --cookies-from-browser vivaldi "https://www.tiktok.com/@hakimdailyyy/video/7588853215685381384"

and it works. Is this problem related to browser impersonation? Why doesn't it work for gallery-dl?

[AlttiRi]

If I'm correct, yt-dlp by default does not use a browser impersonation.

• https://github.com/lexiforest/curl_cffi
• https://github.com/lexiforest/curl-impersonate

You need manually to pass something like --impersonate chrome-110:windows-10.

(gallery-dl does not support it anyway)

[sondhg]

Thanks for the reply. Yes I know that. My takeaway from reading the docs of gallery-dl is that yt-dlp is a necessary dependency for gallery-dl to be able to download videos. I faced the browser impersonation problem before. That's why I installed curl_cffi as dependency of yt-dlp.

Edit: I installed gallery-dl via uv tool install gallery-dl --with 'yt-dlp[default,curl-cffi]'. It has been working fine for months.

After I installed gallery-dl this way, I was able to use it to download TikTok videos for months, even without providing a cookies txt file. Just recently I can't download TikTok videos anymore, even after I provide the cookies txt file. So I'm not sure whether TikTok themselves implemented some security measures against scraping, or this is a problem with gallery-dl

[skymatr1x]

this is an active yt-dlp issue: yt-dlp/yt-dlp#15418

[sondhg]

Seems like it is. I was able to download a TikTok video today, but it's slow. Sometimes it still doesn't work. Let's wait and see if the yt-dlp dev fixes this. Thanks for the response.

[skymatr1x]

since the day that issue was created, it hasn't been working most of the time for most countries.

i haven't had any notable speed issue though.

yeah, hopefully soon.

@CasualYT31 have you come across this issue with tiktok recently? curious whether your PR is able to bypass this.

[CasualYT31]

I haven't come across it. So I'm guessing the regions that I use haven't been impacted yet.

I also highly doubt my PR will work around this considering we've largely copied yt-dlp's approach.

I have some time today to look into this. It probably won't be a simple fix though.

[CasualYT31]

I can replicate the issue using the Australia region, both with the latest version of gallery-dl and with my PR. As mentioned in the yt-dlp thread, it's probably some JavaScript challenge. I've never dealt with these so I don't know how to address the problem yet. I'll keep digging.

[CasualYT31]

I see, the challenge page includes some JSON and JavaScript that would get executed in a normal browser, but obviously it won't work when using the requests library, hence the suggestion of using deno (in the yt-dlp thread) to address the problem. But hopefully we can take a different approach that doesn't require a separate dependency.

For some strange reason, I am able to get the rehydration data using this simple code, even when using a region that breaks yt-dlp and gallery-dl:

from requests import get, Session

r = get("https://www.tiktok.com/@erinemoriarty/video/7580919616520588557")

print(r.text)

# or

s = Session()

r = s.get("https://www.tiktok.com/@erinemoriarty/video/7580919616520588557")

print(r.text)

I am also able to do the same with requests-html, which gallery-dl does not use at this time:

from requests_html import HTMLSession

session = HTMLSession()

r = session.get("https://www.tiktok.com/@erinemoriarty/video/7580919616520588557")

print(r.text)

So yt-dlp and gallery-dl must be making the request in such a way that it triggers the challenge. I wonder if it's just a matter of adjusting the way we make these requests in the extractor 👀?

[mikf]

Maybe it's just due to some HTTP header value like User-Agent.
gallery-dl uses the latest Firefox UA, while requests on its own uses python-requests/2.32.5

[CasualYT31]

You're right, with this code I get the challenge response:

from requests import get

r = get("https://www.tiktok.com/@erinemoriarty/video/7580919616520588557", headers={
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:146.0) Gecko/20100101 Firefox/146.0",
})

print(r.text)

Wonder if it's worth using the python-requests/2.32.5 user agent in the PR for the time being? I've got a feeling this problem is going to come back to bite us though.

[CasualYT31]

Although it might not be the only piece to the puzzle, because using this code in _extract_rehydration_data() does not always work:

response = self.request(url, headers={
    "User-Agent": "python-requests/2.32.5",
})

[CasualYT31]

FWIW I seem to be able to extract the rehydration data consistently without any changes to the code if I use cookies (I extracted these cookies right after loading the video in the browser though so that might be why). And I actually always get 403 forbidden if I try to use the python-requests/2.32.5 user agent when using cookies so that seems like a bad way forward.

[skymatr1x]

thanks for looking into it! sorry i couldn't help with the code, it's out of my scope.

I can replicate the issue using the Australia region

same here, lots of regions get this error, only a small number seem unaffected.

FWIW I seem to be able to extract the rehydration data consistently without any changes to the code if I use cookies (I extracted these cookies right after loading the video in the browser though so that might be why).

rehydration data is different than downloading the files like usual right?

i've tried passing the extracted cookies.txt after loading the video, but that itself didn't help to bypass the error.

[CasualYT31]

For future reference, Germany and Switzerland aren't affected.

rehydration data is different than downloading the files like usual right?

Sorry, I was not clear. The extraction is failing at the rehydration data extraction step as it assumes that it doesn't encounter any JavaScript challenge. So when I say that the rehydration data could be extracted, I mean that the post could be downloaded.

i've tried passing the extracted cookies.txt after loading the video, but that itself didn't help to bypass the error.

I've noticed a few users mentioning this but it wasn't my experience earlier today. Bit of a strange one.

[skymatr1x]

For future reference, Germany and Switzerland aren't affected.

BE, MX, NL, US work for me.