For marketers

Transcribe customer calls for marketers

Transcribe customer calls for marketers: a timecode on every line, the phrase for the landing page found in the text, and a CSV for the sheet. Pro $10 a month.

3 files a day up to 30 minutes, free99+ languagesSpeaker separation and timestamps
In short

Upload the recording of a customer call and the words come back as text with a timecode on every line, so a quote for a landing page can be checked against the audio. One file is up to 30 minutes, and Pro is $10 a month paid yearly.

The case

Why the voice of a customer is raw material

The voice of a customer is raw material, and it is usually thrown away. A call ends, the recording sits in the CRM as an attachment nobody opens, and the phrase that would have sold the page is lost in it. A transcript turns that attachment into text that a writer can quote, a spreadsheet can count and an editor can cut into a script.

The numbers

What a customer call gives the campaign

One call

up to 30 minutes on the free plan, no minute limit on Pro

The pipeline

a REST API and an MCP server with 4 tools; the API plan is billed on the audio you send

An API price elsewhere

AssemblyAI publishes $0.21 an hour for a pre-recorded file (Universal-3.5 Pro) and $0.02 an hour for the speaker separation

How it works

How to transcribe a customer call

  1. 1

    Save the recording of the call

    The file the phone or the meeting tool produced. It may be up to 30 minutes, so a discovery call fits as it is.

  2. 2

    Upload it and split the two voices

    Drop the file on the widget above. Switch on the speaker split, so the words of the customer do not end up in the words of the seller.

  3. 3

    Find the phrase for the page

    Every line carries the time it starts and ends. Open the text and find the words the customer used: the sentence is there with the minute next to it.

  4. 4

    Move the excerpts into the sheet

    CSV gives one row per segment with the time and the speaker. DOCX for the brief on Pro, Markdown for the page draft, TXT for the archive.

Tips

  • Name the file after the company of the customer.

  • Split the voices before you look for a quote.

  • Keep the consent with the transcript, not in a drawer.

Questions this page answers

Can the quotes be used without permission?

That is your call and your lawyer's, not ours. We only turn the recording into text; the consent for a marketing quote is yours to obtain and to keep.

What does an API for this cost?

AssemblyAI publishes $0.21 an hour for a pre-recorded file (Universal-3.5 Pro), and the speaker separation costs $0.02 an hour on top; that is assemblyai.com/pricing, read on 21 October 2026.

How many calls a month does the plan cover?

Pro carries no minute limit of audio a calendar month. The free plan is $0 with 3 files a day and asks for no card.

Is the file sent to the CRM automatically?

Not by us. The API and the MCP server with 4 tools let your own system send the file and take the text back, so the pipeline stays in your hands.

What happens to the voice of the customer?

A Russian or mixed file is processed by our own engine outside the United States. A pure English file may go to a cloud speech provider there when that fallback is on.

Sources

  1. A
    AssemblyAI pricing (the API and the speaker separation)assemblyai.com · checked 21 October 2026

Reading: 5 minutesUpdated Transcribsi editorial team

Drop a recording — the text arrives in a minute

Speakers by name, a time on every line, export to TXT, SRT, VTT, CSV and Markdown.

3 files a day up to 30 minutes, free99+ languagesSpeaker separation and timestamps