How to transcribe a meeting or an interview recording

November 8, 2026 · All posts

A recording of a meeting or an interview is only useful if you can find things in it. A transcript with timecodes lets you search for a name, copy an exact quote and jump to the moment someone made a decision. This guide shows how to get one with Echoteka transcription, how to turn it into a short summary, and what to do before you record so the text comes out clean.

Before you record

• Ask for consent. Tell the participants you are recording and that the recording will be transcribed. In many countries this is a legal requirement, and it is good manners everywhere.
• Record each voice close to a microphone. A laptop on the far end of a table hears the room more than the people. In a call, ask everyone to use a headset.
• Keep music and notifications off. Music under speech makes recognition harder, and a notification chime can cover a word.
• Note the names and terms that will come up: guests, companies, products, jargon. You will paste them into the tool.

Step by step

1. Export the recording. Most call apps and recorders save MP3, M4A or MP4. Phone voice memos are M4A or AMR. All of these are accepted as they are.
2. Open transcription and upload the file. For a video, only the sound is taken.
3. Choose the language or leave it on Auto: English, German, French, Spanish, and Bulgarian in beta.
4. Paste names and terms into "Names and terms". The list works as a hint, so "Ana Petrova" is not written as "an upper over".
5. Press Transcribe. A long file runs in the background. Close the tab if you like: we email you when it is ready, and the transcript stays under My transcripts.

Reading and fixing the transcript

The text comes in paragraphs with timecodes like [12:31]. A player sits next to it: click a word or a timecode and the recording plays from there, with the current word highlighted. Use search to find every mention of a name or a topic, and fix mistakes right on the page. Then copy everything with one button, or download:

From transcript to summary and action items

Right after recognition the app writes three summaries on its own: In brief (three lines with the essence), Topics with timecodes (click a topic to jump to that place in the text and the player) and Action items: who does what by when, only if it was said in the recording. After sign-in, Another report from the text turns the transcript into a short text in one of five formats:

• Meeting: the main points, the decisions, who does what by when, and open questions.
• Interview: the key answers and the quotes worth keeping.
• Lecture: the main ideas in order, like lecture notes.
• Podcast: a description of the episode.
• Your own request: for example "list every number mentioned and what it refers to".

A summary is a starting point, not minutes you can sign without reading. Check it against the transcript, especially names, figures and deadlines.

Interviews: getting quotes right

For a published interview, quote from the transcript and listen to the quote once more in the player before you publish. Recognition can drop small words such as "not", which change the meaning. If the interview was recorded in a noisy cafe or an echoey room, clean it first with the interview set in audio restoration; the clearer the speech, the fewer fixes later.

Limits and privacy

Transcription is free within a daily limit: 30 minutes a day when signed in and 5 minutes without an account, reset at midnight UTC. Without an account the text is shown on screen, and downloads need sign-in. A long meeting that does not fit into today's minutes can be cut into parts with the audio cutter and transcribed over several days.

The audio is deleted right after transcription. The text is kept for 30 days, Delete removes it at once, and recordings are not used to train models.

What the first version does not do

• No speaker labels yet. In a dialogue the text comes as paragraphs split by pauses. You can type names in while editing.
• No bot that joins your calls. Record the call in your call app, then upload the file.
• Overlapping voices may leave gaps in those places.

Making subtitles for a video instead? See how to make SRT subtitles from a video for free. Turning a voice message into text? See voice message to text.