How do you transcribe a lecture or meeting recording with AI?

A lecture or meeting transcript in colloquial language, with speaker names, followed by a summary or booklet.

From a recording to text you can work with

You have a two-hour lecture recording, a meeting, or a live session. You need text to make a summary or booklet, or to search through it. AI transcription now takes minutes and handles Egyptian Arabic well.

Where to upload it

ToolUseful when
Gemini or Google AI StudioA long recording where you want a complete transcript in colloquial language with timestamps.
Gemini NotebookYou want to ask about the recording and return to the relevant place in the text.
ChatGPTA short recording, or one you want to turn straight into a summary.

If the video is large, convert it to audio first. See the method in the sources lecture.

Ask for the transcript properly

Transcribe a recording

Copy the prompt as it is, or tailor it with your details first.

Tailor the prompt to your workOptional: add your details and the prompt adjusts itself
Prompt details
View full prompt

If the recording is longer than an hour, split it into parts and transcribe each part separately.

Review before building on it

Listen to sections from the beginning, middle, and end of the recording and compare them with the text. Focus on names, numbers, and terms, since these are most often transcribed incorrectly. Search the text for “unclear” and return to the timestamp.

After transcription

The text can become a meeting summary with tasks (see the meeting decisions workflow), an educational booklet, or posts through content repurposing.

Try it in your work

Transcribe 10 minutes

Take 10 minutes from a recording you have, transcribe it using the prompt above, and check it by listening line by line.

Questions? Message us