Transcribe a long English recording, all three hours of it

What happens to a three-hour recording

• You upload one file of up to 2 GB; a session saved as several files can go in the same job, joined in name order
• The recording is cut into parts of up to ten minutes, so three hours becomes about eighteen parts
• Each part is transcribed in full with the same instructions; no part is skipped or summarized
• Timestamps are shifted by where each part starts, so a sentence two and a half hours in carries a time such as [2:31:07]
• The parts are joined in order into one transcript you can edit, search and download as Word, PDF, TXT, SRT or VTT

Why a chat assistant struggles with the same file

Chat assistants are good at a short, clear clip: a voicemail, a two-minute memo. Three hours is different. Assistants that accept audio limit how long a file can be, many free plans cannot take audio at all, and three hours of speech is tens of thousands of words, far more than one reply holds. Asked for everything, an assistant tends to shorten the text, slide into a summary or stop partway, and it cannot produce reliable timestamps for a file that long. Splitting the work into parts is what makes the full transcript possible.

Depositions, hearings and legal recordings

Lawyers, paralegals and people representing themselves use this to get a searchable working copy of a deposition video, a hearing, a police interview or an arbitration session. Every paragraph has a timestamp, so you can go straight to the minute a witness said something and play it back. Speaker changes are marked "Speaker 1:", "Speaker 2:" when they are clear; names appear only if they are said in the recording.

Two limits matter here. First, speaker numbers are assigned per part of about ten minutes, so on a three-hour deposition the attorney who is Speaker 1 early on may become Speaker 2 later; read through, fix the labels, then use Find and replace to put in real names. Second, this is a draft made by an AI speech model, not a certified transcript. Where a court requires a certified record from a court reporter, use this to prepare and order the certified version for filing.

Podcasts, board meetings, focus groups and lectures

A long podcast interview, an annual meeting, a two-hour focus group or a lecture series all come out the same way: complete text, timestamped, ready to edit. Focus groups are the hardest of these, because people talk over each other; overlapping words are marked [inaudible] rather than guessed. Accents from Texas to Glasgow, Lagos, Mumbai or Brisbane are written as the English that was spoken. Strong accents, distant microphones and noisy rooms all cause mistakes, so check names, figures and anything you will quote.

If you need more than the text, other tools here run the same full transcription and add to it: Meeting Recording to Minutes adds decisions and action items, Podcast Transcript and Show Notes adds chapters and quotes, Interview Transcription labels every paragraph with a speaker, and Lecture Recording to Notes adds study notes.

The limits, stated plainly

• One payment of $5: one recording of up to 6 hours
• Unlimited, $19 a week (renews weekly, cancel any time): up to 12 hours per job, every tool
• Longer than that: split the file and run the pieces as separate jobs
• File size: up to 2 GB per upload; video files are fine, only the sound is used
• Optional translation into Spanish or another listed language, printed under each English paragraph

How the tool works, what it costs and how your files are handled is on the Long Audio & Video to Text page; prices for every tool are on the pricing page.

Last updated 25 October 2026