All posts
Guides5 min read

Turn old voice memos and interview files into named notes

A recording from last year can still tell you who said what, once the voices in it are known.

Siya RagadeGuides

To turn old voice memos and interview files into named notes, import the audio into a notetaker that recognizes voices, name each new person once, and the notes come back with names instead of Speaker 1 and Speaker 2. Dragon Notes does this on iPhone, Android, Mac and Windows, for recordings up to four hours long.

Most of us have a folder of recordings we never went back to. A journalist has thirty interviews from last spring, a founder has customer calls saved as voice memos, and a student has a semester of lectures sitting on a phone. The quote you remember is in there somewhere, but finding it usually means scrubbing back and forth through forty minutes of audio.

Getting a voice memo to text is the easy part now. Getting the names right is the part most tools skip. With a live meeting, a tool can at least guess names from the guest list, though calendar invites can’t tell you who’s talking even then. An old recording does not come with a guest list at all.

How to transcribe voice memos

For a quick voice memo transcription, the built-in option on iPhone or Mac is hard to beat. Apple explains that Voice Memos can turn speech in a recording into text, and that transcripts can be copied and searched. Its guide says nothing about telling speakers apart, so a two-person conversation comes back as one long block. That is fine for a note to yourself. For an interview, where the whole point is knowing which lines were the question and which were the answer, it falls short.

Splitting a recording by voice is the first step in how AI notetakers tell speakers apart, and putting a real name on each voice is the second. Plenty of transcription tools stop after the first step, which is why your transcript says “Speaker 2” on so many imported files. In interview transcription, the names are arguably the most valuable part. A great quote is not much use if nobody can say who said it.

What happens to names in an imported recording?

Imported files are where notetakers differ the most, and it is easy to miss when they are only compared on live meetings. A recording from last spring has no calendar invite or video call to borrow names from, so the tool has to go on the voices alone. Some tools learn a voice once it is tagged, and Otter says tagging a speaker teaches it to recognize that voice in future conversations. Others keep names to one file. Plaud, for example, says names added for other people apply only to that recording.

That difference adds up fast. Picture importing twenty interviews where the same editor asks the questions in every one. One tool asks for her name once, and the other asks twenty times.

How to turn voice memos into named notes with Dragon Notes

Dragon Notes imports recordings on iPhone, Android, Mac and Windows, and it only takes a few minutes:

  1. First, find the file. To transcribe voice memos from an iPhone, share the memo or save it to Files, and on Mac, drag it out of the Voice Memos app.
  2. Then, import it into Dragon Notes. Common audio and video files work, including .m4a voice memos, .mp3 files and video recordings, up to four hours each.
  3. Once the recording is transcribed, anyone already named in Dragon Notes is recognized by their voice straight away.
  4. New voices are separated and named where Dragon Notes can tell who they are, and any others take one click to name.

From then on, Dragon Notes recognizes those people in every recording, whether it is a file from last spring or a live conversation next week. If a name ever needs changing, click it and pick the right person. To see how the recognition holds up with several people around one phone, read how we tested speaker recognition in a real room.

An interview from last spring. Maya is known by her voice; the new voice takes one click to name.

Which recordings are worth importing first?

Start with the recordings that have the most people and the most at stake. Interviews are the obvious choice, since a quote only counts when it sits with the right person. Client calls come a close second, because that is where promises and feedback get made and then quietly forgotten. Lectures work well too, especially seminars where the questions from the room are worth keeping apart from the lecturer.

Older recordings tend to be messier than live meetings. Think of one phone on a cafe table, a coffee grinder in the background and two people finishing each other’s sentences. Before quoting anyone from a file like that, it is worth knowing why AI notetakers mix people up and how to spot it. For recordings where people talk over each other, read our guide on how to transcribe interviews when people talk over each other.

Frequently asked questions

Can I transcribe voice memos with speaker names?

Yes, as long as the tool both separates the voices and lets you put names on them. Some only do the first part, which leaves you with speaker labels like Speaker 1 and Speaker 2.

Does iPhone Voice Memos identify speakers?

Not according to Apple’s guide, which covers viewing, copying and searching transcripts but makes no mention of speakers. It works well for solo notes, and less well for anything with two or more voices.

How long can an imported recording be in Dragon Notes?

Up to four hours per file, which covers most interviews, lectures and meetings. An all-day workshop can be split into parts first.

What file types can Dragon Notes import?

Common audio and video files. Dragon Notes can transcribe M4A voice memos, MP3 files and video recordings, on iPhone, Android, Mac and Windows.

Start with one recording

It helps to start small: one interview, one client call or one lecture you keep meaning to go back to. Once the people in it are named, every recording after that gets quicker, because Dragon Notes already knows who they are. It is free for the first week, then $1.99 a month. Try it on that first recording, and see who it remembers!

Siya RagadeNov 16, 2026