Last verified: July 2026
NotebookLM Says "Transcript Not Available"? Here's the Real Reason — and the Fix
You pasted a YouTube link into NotebookLM as a source and it failed. Here's why that happens and how to get a usable transcript anyway.
Failed to add source. This video doesn't have a transcript available.
NotebookLM shows an error like this when it cannot find a YouTube caption track to read.
Why NotebookLM can't read this video
NotebookLM doesn't watch or listen to YouTube videos. It reads the caption track attached to the video on YouTube's servers. When that caption track doesn't exist or isn't accessible, there's nothing for it to import — no audio fallback, no transcription of its own.
No captions exist for this video
The uploader never added captions, and YouTube hasn't generated auto-captions for it yet. Without a caption track, NotebookLM has no text to pull in.
Captions are disabled by the uploader
Some creators turn off captions entirely, even when auto-captions would otherwise be available. NotebookLM sees the same "no transcript" state either way.
The upload is too new
YouTube's auto-captioning can take anywhere from a few hours to a couple of days to process after upload. If you're watching something fresh, the caption track may simply not exist yet.
The video is private, unlisted, or age-restricted
NotebookLM (and most tools like it) can only pull from videos it can access as a normal viewer would. Private or restricted videos fail before captions even become relevant.
What each tool does when captions are missing
| Video situation | NotebookLM | ChatGPT | Gemini | YouTLDR |
|---|---|---|---|---|
| No captions at all | Fails to add source | Can't summarize — no transcript to read | Reports no transcript available | Transcribes the audio directly |
| Captions disabled by uploader | Fails to add source | Fails or asks you to paste text | Fails or gives a generic answer | Transcribes the audio directly |
| Auto-captions only in another language | Imports garbled or wrong-language text | Summarizes the wrong language poorly | Same caption-language limitation | Transcribes in the video's actual spoken language |
| Brand-new upload | Fails until YouTube processes captions | Fails until captions exist | Fails until captions exist | Works immediately — no caption wait |
| 3-hour lecture or podcast | Works only if captions exist | Often truncates or times out | Often truncates or times out | Handles up to 5 hours of audio |
| Non-English video | Works only if that language has captions | Limited by available caption language | Limited by available caption language | Supports 125+ spoken languages |
How to get the transcript anyway
Paste the YouTube link into YouTLDR
Copy the same URL you tried in NotebookLM and drop it into YouTLDR's homepage. No caption track required.
YouTLDR transcribes the audio directly
Instead of reading a caption file, YouTLDR listens to the video's actual audio and generates its own transcript — so a missing or disabled caption track doesn't matter.
Get your transcript and TLDR summary with timestamps
Copy the transcript back into NotebookLM as a text source, or just keep working in YouTLDR — export it, search it, or read the summary directly.
Why YouTLDR can read videos NotebookLM can't
NotebookLM, like most AI research tools, imports YouTube videos by reading the caption track — it never actually processes audio. That means its success depends entirely on whether YouTube (or the uploader) happened to generate or enable captions for that specific video.
YouTLDR skips that dependency. It runs the audio itself through speech-to-text transcription (Whisper large-v3 class), so it works whether or not a caption track exists, in any of the 125+ languages it supports, for videos up to 5 hours long.
That also means it handles the edge cases that break caption-based tools: brand-new uploads, creators who disable captions, and non-English audio with no matching caption language. You can try it free, without creating an account, before deciding if you need it regularly.
Frequently asked questions
Is this NotebookLM's fault?
Not exactly. NotebookLM is built to read YouTube's caption data, which is a reasonable design choice — it's just not built to transcribe audio itself. The gap shows up whenever a video's captions are missing, disabled, or in the wrong language.
Will YouTube add captions later so NotebookLM works?
Sometimes. If the upload is just new, auto-captions may appear within a day or two and NotebookLM might pick it up on a retry. But if the uploader disabled captions or none were ever generated, that error won't resolve on its own.
Does this work for live streams?
Only after the stream ends and becomes a regular uploaded video. Live streams in progress don't have a finished audio file to transcribe yet.
How long can the video be?
YouTLDR can transcribe videos up to 5 hours long, which covers most lectures, podcasts, and long-form interviews that trip up caption-dependent tools.
What languages are supported?
YouTLDR supports 125+ languages for transcription, so it can handle videos where YouTube never generated an auto-caption track for that language at all.
Do I still need NotebookLM after this?
You can. Copy the transcript YouTLDR generates into NotebookLM as a text or pasted source — NotebookLM can work with text you provide even when it can't pull the video's captions itself.
Related fixes
NotebookLM Says "Transcript Not Available"? Here's the Real Reason — and the Fix
You pasted a YouTube link into NotebookLM as a source and it failed. Here's why that happens and how to get a usable transcript anyway.
Transcribe this video now