SubGrab

Every Transcript Error, What It Actually Means, And Whether Retrying Helps

Some failures are worth a second try, some need a different link, and some will never work. A plain-English decoder for every message you can see — and what happens to your credits.

A failed transcript is annoying. A failed transcript with a vague message is worse, because you can't tell whether to try again, fix something, or give up.

So here is every failure you can actually see, what causes it, and the only question that matters: does trying again help?

First, the thing that applies to all of them: a failed transcript is always refunded, automatically. You are never charged for a transcript you didn't get. If you see a failure message, the credits are already back on your balance before the message reaches you.

"Please try again in a few minutes"

Retrying helps. Usually immediately.

This covers the genuinely transient failures, and they're more common than you'd think:

  • A platform's anti-bot challenge refused this particular request. TikTok does this constantly — see why TikTok fails then works for the mechanism.
  • A platform's API was slow enough to time out on one call.
  • A rate limit, hit because a platform saw too many requests in a short window.

We already retry these automatically before showing you anything, so if you're seeing this message we've tried more than once — and we keep going. SubGrab retries the job for you 10 minutes, 40 minutes and then 2 hours after each failure (up to three more tries), charges you only if one works, and emails you when the transcript is ready. You can stop that from the failed page, or try again yourself: these clear on their own.

"This video requires a login" / "only shown to signed-in viewers"

Retrying will not help. The access has to change — or you supply the media.

The video is private, members-only, subscriber-only, age-restricted to signed-in accounts, or limited to certain accounts. Transcription works on what's publicly viewable, and this isn't that. (A Vimeo video behind a password is different: SubGrab asks you for the password before charging anything, uses it for that one job and doesn't store it. That includes a password-protected Vimeo showcase.)

When the platform gives its own reason, the message quotes it. SOOP, for example, answers a 19+ broadcast with "Only users older than 19 are able to watch this video." Since 19 November 2026 these walls are recognised on every platform and turned away before any credit is taken; before that, one that had no message of its own read "AI transcription failed. Please try again."

The reliable test: open the link in a private browsing window while signed out. If it doesn't play for you there, it won't work here either.

What does work: if you have access to the video yourself, download it and upload the file directly. That sidesteps every privacy setting, because you're supplying the media rather than asking the platform for it. If you can play it but not download it, the failed page offers Record it from a tab instead on desktop Chrome or Edge: play the video in another tab, share that tab with its audio, and its sound is recorded as it plays (up to 60 minutes a recording) and transcribed like an upload.

"This Google Drive file has hit Google's viewing limit"

Retrying won't help today. It will help tomorrow.

Google caps how often a Drive file can be watched by signed-out viewers. The limit belongs to the file, so it isn't about your link, your account, or the tool. Make a copy of the file for a fresh limit, or upload the video directly. You can also try Choose it from Google Drive on the failed page, though whether it gets past this limit hasn't been measured yet. Full explanation in what "quota exceeded" actually means.

"Set sharing to Anyone with the link"

Retrying won't help. One setting will — or, on Google Drive, one choice.

Different from the quota message above, and the fix is exactly what it says. In Drive or Dropbox, open the share settings and make sure the link is readable by anyone who has it, not just people you've invited.

For a Google Drive file there's a second way that leaves the sharing alone: the failed page offers Choose it from Google Drive, which opens Google's own file picker. Pick the file and SubGrab can open that one file only; the permission ends within an hour.

"This VOD is subscriber-only"

Retrying will never help.

A Twitch VOD behind a channel subscription is not reachable without that subscription. Try a public VOD or a clip from the same channel instead.

"This video is unavailable"

Retrying won't help.

Deleted, removed by the platform, or the link points at something that no longer exists. Double-check you copied the whole URL — a truncated link can produce this too.

"This video isn't available in the country our servers are in"

Retrying won't help.

The uploader or the platform limits which countries can watch it, and our servers' country isn't one of them. When the platform names the country, the message does too, for example "only shown to viewers in Japan". If the video plays for you, upload the file instead, or record it from a browser tab in Chrome or Edge on a computer.

"This video doesn't have an audio track"

Retrying won't help, and nothing is wrong.

Exactly what it says: a silent video, or a post that turned out to be images rather than video. There's no speech to transcribe. Common with TikTok photo carousels and with X/Twitter links that point at a text post.

"This video's audio format couldn't be processed"

Retrying rarely helps.

Usually DRM-protected content whose audio can't be read. A stream we can tell is still live gets its own message instead — "This stream is still live" — before anything is charged. Live streams work once they've finished and been published as a recording.

"We couldn't read this video's page — please report this link"

Retrying might help, but please do report it.

This is the one message that means we have a problem, not you. It means the platform changed something about how its pages work and our extraction hasn't caught up. It's the only message on this page that's worth sending to us, and it's deliberately narrow so that it stays meaningful.

We used to show this for things that weren't our fault at all — Google Drive quota limits, slow API calls, TikTok's challenge wall. That made it useless as a signal, for you and for us. Now it means what it says.

"We couldn't get this video. The platform says: …"

Maybe. Read the platform's sentence.

This is a failure no rule of ours recognises yet, so SubGrab shows the platform's own explanation instead of guessing. If it reads like a temporary problem, try again; if it reads like a restriction ("for members only", "not available to you"), retrying won't change it. Every one of these alerts us, and each gets its own plain-English message once we have seen it.

No speech detected

Retrying won't help.

The audio was processed successfully and contained no recognisable speech — music-only, ambient noise, or silence.

Turned away before you're charged

Some links are refused before any credit is taken, because we can tell up front that they can't work. Nothing is charged, so nothing needs refunding:

  • "Videos over … hours are not supported" — the limit is 10 hours, and 16 for YouTube and Rumble videos and Kick, SOOP and CHZZK streams. Save the part you need and upload it.
  • "This stream is still live" — paste it again once the broadcast ends and the recording is published.
  • A LinkedIn Live recording — LinkedIn doesn't publish Live recordings on the public post page. Upload the recording if you have it.
  • An X Space the host didn't record — X has no replay to transcribe. Recorded Spaces work.
  • A Reddit post that links to a video on another site — paste the link the post points to instead.
  • A Telegram channel that keeps its videos in the app, or a video too large for Telegram's web preview — if you can save the video in Telegram, upload the file.
  • A Facebook video we can't reach — Facebook hides some videos from signed-out visitors; see the raw-string note at the end of this page.

The one-line summary

If the message suggests trying again, try again — or let the automatic retries do it. If it names something about the video — private, deleted, region-locked, silent, subscriber-only — that's a property of the video and no retry changes it. If it asks you to report the link, that one's ours.

And in every case, the credits are already back.

Related Reading

Try a transcript →

The raw strings, if you are looking at another tool's output

Our messages are rewritten for humans. Most transcript tools surface the

underlying extractor output instead, and if you are staring at one of those,

here is what it actually means. These are verbatim strings taken from our own

production logs, not paraphrases.

ERROR: [facebook] <id>: Cannot parse data — reads like a broken tool and

is not one. Facebook served a logged-out client a decoy page instead of an

error: its "Discover popular videos" hub, with no playable URL anywhere in it.

The extractor had nothing to parse and said so. Facebook serves the same decoy

for a deleted or mistyped video, so the string alone doesn't say why.

Retrying the same request will not help. Friends-only, group and private

posts end here for every tool; some public videos Facebook hides from signed-out

visitors can still be reached, so on SubGrab it is worth pasting the link — one

we can't reach is refused before you're charged. It is the single largest

failure class in this product.

ERROR: [Rumble] <slug>: Unable to download webpage: HTTP Error 403: Forbidden

— a rate wall, not a permission problem. The tell is that it is not stable: the

same link answers, then refuses, then answers again. We reproduced exactly that

on one link in three consecutive requests. Retrying does help, and often

immediately.

ERROR: [GoogleDrive] <id>: Unable to download JSON metadata: HTTP Error 403: Forbidden

— almost always a sharing setting rather than anything technical. The file is

shared with specific people rather than with anyone holding the link.

Retrying will not help, but you can fix it in about ten seconds if you own

the file.

ERROR: [vimeo] <id>: This video is protected by a password, use the --video-password option

— exactly what it says. On Vimeo, SubGrab asks you for the password before taking

a credit (how); elsewhere somebody has to remove the password, or you

download the video yourself and upload the file.

ERROR: [TikTok] <id>: Video not available, status code 10231 — TikTok's own

API refusing a video, with a numeric code it does not document, and its cousin

10204 ("Your IP address is blocked from accessing this post"). We first measured

10231 as per-video: the failing video returned it on three of three attempts

while a known-good control resolved three of three in the same minutes. But

every one of those attempts came from one country. Asked again through a US

connection, both refused videos played 12 of 12 times — so it is a refusal to a

country, and SubGrab now retries these from the United States.

Retrying from the same place will not help; from another country it can.

A video TikTok refuses everywhere, such as a paid drama episode, stays refused.

ERROR: [wetv:episode] <id>: This video is not available from your location due to geo restriction

— honest and final. No tool gets past this, whatever it claims.

An extractor error has occurred. (caused by KeyError('sourceURL')) — a bug

in the extractor itself, not in your link. This one is a bare dictionary lookup

on a field the DASH specification marks optional. Retrying will not help

until the tool is fixed upstream.

WARNING: Extractor <name> returned nothing followed by no error at all —

the extractor ran, found no video, and exited successfully. Treating that exit

code as success is how a tool ends up handing you an empty file instead of a

failure. Retrying will not help; if you can play the video yourself, save it

and upload the file.

Written by the SubGrab team. About SubGrab says who we are and where your audio goes; questions and corrections go to help@subgrab.com.