You Have the Recording but No Transcript. What Now?
Someone recorded it, nobody transcribed it, and now you need to find one sentence in ninety minutes of audio. Here is how to turn the file you already have into text you can search.
This is the most ordinary transcription problem there is. A meeting happened, someone recorded it, the platform did not produce a transcript, and now the information is locked inside ninety minutes of audio that nobody will ever listen to again.
Why the recording exists without a transcript
The causes differ per platform and they all end in the same place:
- Zoom only creates a transcript when the host recorded to the cloud with the Audio Transcript toggle already on, which requires a paid plan.
- Webex only transcribes cloud recordings. A local recording produces a video file and nothing more.
- Microsoft Teams ties transcription to admin policy, licence and role, so a recorded meeting can easily have no transcript attached.
- Slack huddles on Free and Pro keep nothing at all, and Business+ recordings arrive as an M4A clip with no text.
- A phone, a dictaphone or a camera never had a transcription step to begin with.
None of these can be repaired retroactively inside the original platform. The transcript was either created at the time or it was not.
Turning the file into text
The file is the only thing that matters now, and a file can be transcribed at any point. What to check before you start:
Format. Kalima accepts mp3, wav, m4a, mp4, mov, aac, flac, ogg, opus, webm, aiff and amr. Video files are fine: only the audio track is used, so an mp4 screen recording works exactly like an audio file.
Length. Uploads are capped per plan: 30 minutes per file on Free, three hours on Plus, five hours on Pro and Team. A recording longer than the cap has to be split, which any audio editor does in a minute.
File size. 100 MB on Free, 500 MB on Plus, 1 GB on Pro and Team. A ninety-minute meeting recorded as m4a is usually well under 100 MB; a video file of the same meeting can be ten times that, so extracting the audio track first is often the difference between fitting and not.
How many. Five uploads a week on Free, 25 on Plus, 50 on Pro and Team.
What you get back
The transcript separates speakers, so a four-person meeting reads as a conversation rather than a wall of text. It is searchable, editable and exports to SRT, VTT, JSON, TXT and PDF. A structured summary and chat over the transcript are available on top of it, which for a long recording is usually the faster way in: ask what was decided rather than reading ninety minutes of text.
All 60 transcription languages apply to uploaded files, and a recording with more than one language in it is handled in the same session.
Doing it live next time
Uploading afterwards works, and it is a repair. The version that does not need repairing is capturing the audio while the meeting happens, which removes the dependency on somebody else's recording settings entirely.
Getting a transcript on your own device
Kalima transcribes the audio your computer is already playing. It does not connect to the meeting platform, needs no licence on anyone's tenant, and adds nothing to the participant list, so none of the causes above apply to it.
- Install the desktop app for Windows or macOS and sign in.
- Start a session and choose the audio: the microphone, the system audio, or both mixed together. For a meeting you want both, so you capture the room and the people on the call.
- The transcript builds while people speak, with speakers separated. If more than one language is in the room, live translation runs in the same session across 60 languages.
- Afterwards you can edit the text, get a structured summary and export to SRT, VTT, JSON, TXT or PDF.
The free plan covers 120 minutes a week with sessions up to an hour, and desktop audio capture and meeting detection are included in it.
Questions people ask
Can I upload a video file? Yes. mp4, mov and webm are accepted and only the audio is used.
What if my recording is longer than the upload limit? Split it. Any audio editor exports a long file into parts, and each part transcribes separately.
How accurate is it on a bad recording? Accuracy follows the audio. A recording made through a laptop speaker in a large room is harder than a headset recording, in the same way it is harder for a person. Adding context terms, such as product names and the names of the people speaking, measurably helps.
Is the audio kept? The recording stays with the session so the transcript can be checked against it, and storage runs from 5 GB on the free plan to 100 GB on Pro.
Try Kalima on your next call.
Live transcription and translation, no bot in the meeting. Free to start.
Keep reading.
How to Get a Transcript of a Slack Huddle
Huddles are the fastest way to talk in Slack and the easiest thing to lose. What Slack keeps depends on the plan and on an admin toggle, and on the lower tiers it keeps nothing at all.
Teams Live Captions Not Working? Every Cause and a Way Around It
Captions greyed out, no Translate option, the wrong language, or nothing appearing at all. Each of those has a different cause in Teams, and most of them are settings a participant cannot reach. This guide names them and shows a way to get readable, saved text regardless.
Webex Transcription Not Available? What Is Actually Blocking It
The transcript option is greyed out, or the meeting ended and the transcript never arrived. In Webex that is usually one of four things, and three of them are decided by someone in Control Hub. This guide tells them apart.