How to record and transcribe an in-person meeting
There are three ways to capture an in-person meeting: You can take notes by hand, record the audio and then send the file to a transcription service, or you can run an AI notetaker that records and transcribes all at once.
What makes an in-person meeting harder to transcribe
Recording a room is different from transcribing a video call. Instead of several people speaking into their own microphones, the transcription software may be working from one microphone picking up several voices at different distances, along with everything else happening in the room. Those conditions put more pressure on both the recording setup and the transcription software.
In practice, a room creates four problems: capturing clear audio from everyone, understanding words through distance, noise, and crosstalk, spelling names and terms correctly, and telling one speaker from another.
Ways to record and transcribe an in-person meeting
There are three practical routes, and the right one depends on how often you do this and what you need afterward.
Take notes by hand
Still the cheapest option, and still the worst for exact wording. Handwritten notes work for short internal meetings where decisions matter more than wording, but they leave you with no exact quotes and no record to check against.
Record the audio, then transcribe the file
This route has two separate steps: capture the audio, then send the file to a transcription service.
You can record on almost any device you already have:
- A phone, using a built-in app such as Voice Memos on iPhone or a recorder app on Android.
- A dedicated voice recorder.
- A laptop, using its built-in recording app, such as Voice Memos or QuickTime Player on a Mac, or Sound Recorder on Windows.
These apps only capture audio. When the meeting ends, you have a file, not a transcript, and you still need to upload it to a transcription service. This works for one-off meetings, or rooms where no laptop is present. The tradeoff is the extra upload step, and no speaker names unless the service adds them.
Use an AI notetaker on the laptop in the room
An AI notetaker records and transcribes in one step, so there is no file to move and no upload. You finish the meeting with a transcript, and often a summary, instead of an audio file. AI allows for greater accuracy and organization than what was possible even a few years ago.
Not every AI notetaker works in a room, though. Many join video calls as a meeting bot, which means they need a call to join. A bot-free notetaker records on the device itself, which allows it to work in a physical room where there is no call to join. This is the route most people end up on once in-person meetings become a weekly thing rather than a one-off.
Which method fits your meeting
Pick by how often you do this and what you need out of it.
Whichever method you pick, always tell everyone in the room that you are recording.
How to transcribe an in-person meeting with Wispr Flow Notetaker
A worthwhile AI notetaker transcribes accurately, identifies speakers, handles names and terminology, and leaves you with a useful meeting record afterward. Wispr Flow Notetaker is built for that.
Wispr Flow Notetaker is bot-free, so it works at a table where there is no call to join. It runs on Mac and Windows and transcribes in 21 languages. In-person meetings and remote calls live in the same searchable meeting history instead of being split across different systems.
- Add the meeting and attendees to your calendar when possible. This gives Wispr names to work with later.
- Add important names, client terms, and acronyms to your dictionary.
- Place the laptop where it can clearly hear everyone in the room, roughly in the middle of the group.
- Tell everyone you'll be recording before you start.
- Start the recording and let the meeting run.
- Review the final transcript and correct any speakers it couldn't identify. Name that person once, and the correction applies across the transcript.
Wispr creates a live transcript during the meeting, then refines that transcript again afterward with more context available. That second pass gives the final transcript another opportunity to improve wording and speaker labels.
Wispr Flow is free to try, and the free plan includes both Dictation and Notetaker, with no credit card required. That means the same product can handle your meetings and the voice-to-text work you do everywhere else.
How to get speaker names right in a room
Getting the words right is only half the job. A quote attributed to the wrong person is worse than no quote.
Video calls have an advantage because each participant usually joins under a name and speaks through an individual audio stream. Around a table, several people may be sharing the same microphone. Plus, with in-person meetings in particular, not every attendee may be listed on the calendar invite.
Solution: Choose an AI notetaker that can pull context from the in-person conversation. Wispr Flow Notetaker builds that context in layers:
- Calendar invite: It starts with the names of the people expected at the meeting.
- Dictionary: Names you have already added give it another source of spelling and identity context.
- Conversation: When people address one another by name, Wispr can use those exchanges to help identify speakers.
- Your correction: If someone still cannot be identified, name that person once after the meeting and apply the correction across the transcript.
There are still limits. If several people are sharing one room microphone and nobody ever says their names, Wispr has fewer clues to work with than it would on a video call where every participant arrives with an identity attached. One-on-ones that are already on your calendar are the easiest case because the attendee list narrows down who can be speaking. For more tips, read our guide to How to know who said what in a meeting.
How to improve in-person meeting transcription accuracy
Most of these fixes apply to any recording method. Where the transcription software makes a difference, we note what to look for.
Capturing clear audio from everyone in the room
The first problem is physical: everyone needs to be heard. A laptop sitting beside one person may pick up that voice clearly while struggling with someone at the opposite end of the table.
Solution: Position the recorder where it can hear the whole group. Put the laptop, phone, or recorder near the middle of the conversation rather than wherever the person taking notes happens to be sitting. Keep papers or bags away from the microphones, and move the recorder away from obvious noise sources when you can.
For a difficult room:
- Test a large room before an important meeting rather than discovering afterward that one side of the table was barely audible.
- Consider a conference or external microphone for unusually long tables or larger rooms.
A better microphone can give the transcription software cleaner audio to work with, but it cannot solve every problem that happens after the sound is captured.
Understanding words through distance, noise, and people talking over one another
People speak at different volumes, accents vary, sentences trail off, and two people may start talking at the same time. Research on in-person meeting transcription describes noise, room reverberation, and overlapping speech as some of the harder conditions for automatic speech recognition.
You can help on your side:
- Cut background noise that is easy to control.
- Avoid prolonged crosstalk when someone is making an important point.
- If a decision gets buried under several people speaking at once, restate it clearly before moving on.
Solution: Choose an AI notetaker built first and foremost around transcription accuracy. Wispr Flow first launched as a voice dictation tool designed to be used on the go in a variety of noisy environments. This heightened precision and the tool's built-in corrections have a direct impact on the correctness of in-person meeting transcriptions.
Transcribing names, product terms, and acronyms correctly
An unfamiliar surname, client name, product term, or acronym can be perfectly audible and still be transcribed incorrectly because the software has never seen the spelling before.
Solution: Use a notetaker that lets you teach it your vocabulary. Wispr Flow Notetaker uses your personal dictionary, so you can add unusual names, proprietary terms, and acronyms in advance instead of asking the software to infer their spelling entirely from room audio.
Correct spellings also make the transcript easier to search and keep the same name from appearing several different ways across one conversation.
How to record and transcribe a hybrid meeting
For a hybrid meeting, the cleanest setup is to have the remote participants join through the same laptop running Wispr Flow Notetaker. That gives the notetaker direct access to the remote audio while the laptop microphone captures the people in the room.
If someone is calling in through a phone line or that setup is not possible, the next-best option is to put the call on speaker and position the phone close enough to the recording device that the remote participant comes through clearly without drowning out the room. Keep the calendar invite updated with everyone who is participating so Wispr has as much speaker context as possible.
When to record, and extra safeguards for high-stakes meetings
Worth recording: customer conversations, user interviews, hiring panels, anything with a decision you will be asked to justify later, and any meeting where you will need someone's exact words. Recurring in-person team meetings pay off too, because the transcript gives you a record of what was decided rather than four people's differing memories.
Not worth it: a two-minute hallway conversation, a meeting where the room is loud enough that you would spend longer fixing the transcript than writing notes, and anything a participant is uncomfortable having on record.
Even with a strong recording setup, an important meeting may deserve more than one layer of protection. If a number, quote, commitment, or attribution could matter later, these extra safeguards make it easier to catch a mistake before you rely on the transcript.
- Confirm consent before you start. Wispr Flow's Notetaker privacy and security guidance makes clear that bot-free recording does not notify everyone or obtain consent for you. Tell participants they're being recorded and follow the requirements that apply to your organization and location. If someone says no, take notes instead.
- Have more than one person record the meeting. For a particularly important conversation, two team members can run their own AI notetakers from different positions in the room. Wispr Flow Notetaker is included in Flow's free plan, while paid team plans make it practical to manage multiple users through centralized billing and team controls.
- Repeat important numbers clearly. If someone gives a price, percentage, date, deadline, or other number that matters, repeat it once before moving on. That gives the transcription system a second, cleaner opportunity to capture it correctly.
- Use a tool that keeps the source audio available for verification. A transcript is faster to search and review, but the original audio gives you something to check when a name, number, or quote looks ambiguous. For high-stakes meetings, that extra verification layer can matter.
- Review the transcript while the meeting is still fresh. Check speaker labels, names, numbers, and major decisions soon afterward, when you can still spot something that doesn't look right without reconstructing the entire conversation from memory.
Start with one meeting in a room
If your week mixes rooms and calls, put the laptop on the table and record. Start with one in-person meeting this week, then read the final transcript. Did the names land? Did your terms come through the way you write them? Would you quote from it?
Start transcribing your meeting in seconds
Wispr Flow meeting transcription FAQs
Where should I put my laptop or phone during an in-person meeting?
On the table, near the middle of the group, with nothing on top of it. Keep it away from projectors, fans, and open doors. Put the meeting on the calendar too, so there are names to work with afterward.
Can I use Wispr Flow to record and transcribe meetings?
Yes. Wispr Flow Notetaker records microphone and system audio on Mac and Windows, produces a live transcript during the meeting, and then processes its own recording into a refined transcript afterward. That setup works for remote, hybrid, and in-person meetings as long as the relevant audio is reaching the computer running Notetaker.
For setup details, read Wispr's guide to recording a meeting with Notetaker.
Can Wispr Flow transcribe an audio file I already recorded?
No. Wispr Flow currently transcribes live speech and meetings it records itself; it does not let you upload an existing MP3, M4A, Voice Memo, meeting recording, or other audio or video file for transcription. It can import certain existing meeting transcripts, but those imports do not cause the associated audio to be transcribed.
See how audio-file transcription works in Wispr Flow for the current supported workflow.
Can Wispr Flow Notetaker record meetings that occur over the phone?
Yes, when the phone conversation's audio is available to the computer running Wispr Flow Notetaker. A call taking place through the computer can be captured through microphone and system audio; if a participant is calling through a separate phone, the practical alternative is to put that call on speaker so the recording device can hear it clearly. Wispr Flow Notetaker does not independently dial into an ordinary phone call as a meeting bot.
For more on how Notetaker handles audio, see Wispr's recording and audio troubleshooting guide.
What happens if I forgot to start recording before the meeting began?
Start recording as soon as you remember. You lose only what happened before the recording started, and the rest of the meeting can still be transcribed normally.
If the decision happened in the first two minutes, say it again on the record before moving on.
What if my in-person meeting isn't on my calendar?
You can still record an unscheduled meeting. The calendar gives Wispr additional context about who is expected to be there, so without it you may need to review the speaker labels and name anyone Notetaker could not identify afterward.
Does Wispr Flow Notetaker keep the audio from my meetings?
Not indefinitely. Wispr's current privacy documentation says uploaded meeting audio is deleted after processing succeeds or permanently fails, while local meeting-audio copies expire after seven days by default; account or organization settings can differ. The transcript itself can remain available until you delete it or a retention policy removes it. Wispr Flow Notetaker does not create voiceprints or biometric profiles.
Read Wispr's Notetaker privacy and security overview for current retention details.
