Can AI transcribe audio? Five assistants, from their makers' pages

Yes, some can, each within limits its maker states. By their makers' own pages: ChatGPT transcribes an audio file you attach on a paid plan, 6 minutes a recording with the model OpenAI names for it; the Gemini app takes a total of 10 minutes of audio free; NotebookLM transcribes an audio file into a source when you import it; Copilot transcribes uploaded recordings through Transcribe in Word; and Claude's list of files you can upload names no audio format. None of those pages describes a subtitle file coming back. We make Burncaps, a transcription tool, and Anthropic, which makes Claude, is one of our suppliers, so read this as a toolmaker's page. Each row below links our page on that assistant, with the date we read its maker's pages.

or drop your video or podcast here
MP4 or MOV, or audio for subtitles.

The video never leaves your browser. On a computer, the sound is heard on your own device unless you choose our speech service or the computer is too slow to listen well; on a phone, it goes to our speech service unless you keep it on the phone. The service gets short pieces of sound, never the picture.

  • Free, no sign-up
  • No watermark
  • Video never uploaded
  • Makers' pages, each dated

Live preview, rendered on this page by the same engine that captions your video.

Five AI assistants and an audio file, side by side

AssistantAn audio file?Limits its pages stateTimes or subtitle filesRead on
ChatGPTYes, on paid plans: attach it and ask for a transcript. Not on Free6 minutes a recording with GPT-6 at Instant, the model OpenAI names; 512 MB; a video file is not accepted as an audio uploadNot stated on the help page15 Nov 2026
ClaudeNo audio or video format on Anthropic's upload list. Voice mode turns your own speech into text as you talk500MB a file and 20 files in a chat, for the files it listsNot described16 Nov 2026
GeminiYes: attach it in the Gemini app and ask about it; the developer page lists transcriptionA total of 10 minutes of audio free, 3 hours with Google AI Pro or Ultra, within rolling limitsThe app page describes answers and summaries, not a subtitle file; the developer page lists timestamps17 Nov 2026
NotebookLMYes: transcribed when you import it, and saved as a source200MB or 500,000 words a source; 50 sources on FreeGetting the transcript out as a file is not described17 Nov 2026
CopilotNot in Copilot Chat's file list. Yes through Transcribe in Word: .wav, .mp4, .m4a or .mp3300 minutes of uploaded audio a month with Microsoft 365, 30,000 with a Copilot licenceSpeakers labelled; saved as a Word document. Subtitle files not described16 Nov 2026

Each row repeats only what our page on that assistant states from its maker's own pages, read on the date shown; follow the link for the sources and the exact words. A maker's page is the authority if it has changed since. We did not test any assistant's accuracy, so this page says nothing about it.

Which one fits what you need

You needThe fit
A summary, answers to questions, notes or an email from a short recordingAn assistant: that is what they do once they have the words
The words themselves, every one, to keep or quoteA transcription tool, then the assistant if you want a summary
Subtitles, or a time on every lineA transcription tool: none of the five assistants' pages describes a subtitle file coming back
A long recording, or one you would rather not send to an AI companyBurncaps: up to three hours and 10 GB, and the file stays in your browser

The route that works with all five: the transcript first

ChatGPT, Claude, NotebookLM and Copilot each list text files among what they read, and the Gemini app's help page says it supports most file types. So the dependable route is two steps.

  1. Drop the recording on the box at the top of this page: an MP3, M4A or WAV, or an MP4 or MOV video.
  2. Correct the script. Click a wrong word and type it right; one click changes it everywhere, which matters for names, because a summary repeats whatever the transcript says.
  3. Press Text, with times if you tick them, or SRT or VTT for subtitles.
  4. Attach the text file to ChatGPT, Claude, Gemini, NotebookLM or Copilot and ask for what you need.

What it costs

Free: 2 videos every 30 days, each with captions and up to 10 clips, no watermark, no sign-up. Then try Pro free for 7 days, no card. Pro is $29 a month for unlimited clips and captions; Starter is $11.98 a month for 6 videos and up to 50 clips. Text and subtitle files are free without limit for recordings up to ten minutes, with no sign-up.

Questions, answered plainly

Can AI transcribe audio?

Yes, within limits each maker states. ChatGPT transcribes an attached audio file on paid plans, the Gemini app takes a total of 10 minutes of audio free, NotebookLM transcribes audio into a source, and Copilot transcribes through Transcribe in Word. Anthropic's upload list for Claude names no audio format.

Which AI can transcribe audio for free?

By the makers' pages: the Gemini app, a total of 10 minutes of audio on its free plan; NotebookLM, with 50 sources on Free. ChatGPT's audio uploads are not available on Free. Burncaps writes text and subtitle files free for recordings up to ten minutes.

Can ChatGPT transcribe audio?

On a paid plan, yes: OpenAI's help page says you can attach an audio file and ask for a transcript, 6 minutes a recording with the model it names. Read on 15 November 2026.

Can Claude transcribe audio?

Not from a file, by Anthropic's own list of files you can upload, which names no audio format. Voice mode turns your own speech into text as you talk. Read on 16 November 2026.

Can AI make subtitles from audio?

None of the five assistants' pages we read describes a subtitle file coming back. Burncaps saves SRT, VTT and ASS files timed to the voice.

Is AI transcription accurate?

We did not test the assistants. OpenAI's page says transcripts may contain errors. Whatever writes it, read the names and numbers back against the recording.

Is a transcript free in Burncaps?

Text and subtitle files are free without limit for recordings up to ten minutes, with no sign-up. A recording over ten minutes gives its first ten minutes free, and all of it as one of your free videos (2 every 30 days), in the 7-day Pro trial (free, no card) or on a paid plan.

Is my video uploaded?

No. Burncaps opens the video in your browser and never receives it. On a computer, the sound is heard on your own device unless you choose our speech service or the computer is too slow to listen well; on a phone it goes to our speech service unless you keep it on the phone (short pieces of sound, never the picture, to Microsoft's speech recognition in the EU, which by Microsoft's terms for the service stores none of it).

Burncaps facts, checked 17 November 2026

What it is: a browser tool that burns animated word-by-word captions into a talking video, or finds the best moments of a talk, podcast or interview and gives them back as ranked, captioned 9:16 clips.

Price: free for 2 videos every 30 days, each one recording with its captioned video and up to 10 clips, no watermark and no sign-up; then the free 7-day Pro trial, no card (your email only); then Pro, $29 a month, gives unlimited clips and captioned videos, or Starter, $11.98 a month, gives 6 videos every 30 days, each fully captioned, and up to 50 clips.

Privacy: the video never leaves your browser. The sound goes only to our speech service, on a phone unless you switch it off, on a computer if you choose it or if it is too slow to listen well. Because the video never reaches us, we cannot keep it or train AI on it, and we train no AI on anything you make.

Files: MP4 or MOV in, MP4 out. Captions on the whole video: up to ten minutes on Free, up to three hours on a paid plan, on a phone or a computer. On iPhones, a whole video over ten minutes has been tested in the iPhone simulator on a Mac (16 minutes at 1080p), not yet on a real iPhone. Clips: up to three hours, on a phone or a computer. SRT, VTT, ASS and text files are free without limit for recordings up to ten minutes; a longer recording, up to three hours, gives its script and those files (the first ten minutes free, all of it with one of your free videos, in the trial or on a paid plan).

Clips: about one clip for every two minutes of talk, up to 30, ranked, the ones we would post first circled, each showing the words it opens and ends on and the reason it was chosen; each cut to 9:16 with ums and dead air removed, an opening card and captions, framed on whoever is talking. Every clip of the whole recording is found and shown, free and with no sign-up; each free video exports up to 10 of them; Starter exports up to 50 clips every 30 days from its 6 videos, and the 7-day trial and Pro export every clip, unlimited.

Languages and styles: captions in the language spoken, twelve Indian languages heard on your device, translation into more than 60 languages, 60 caption styles.

After it listens: correct any word (one click changes it everywhere), add or cut words, one to six words at a time, position, size, text and highlight colours, and names this browser remembers, all free and without listening again.

Made by: Jordan Labs Ltd, UK company no. 17358646.

Against other tools: sixteen caption tools compared, from the vendors' own pages.