How to Transcribe a Zoom Meeting

If you need to transcribe Zoom meetings — for searchable notes, accurate minutes, an accessibility record, or a follow-up email you can quote from — the cleanest approach is to record the call to your own computer and transcribe that local file. This guide covers how to do it accurately and, crucially, privately: your recording never has to leave your PC.

Why transcribe the local recording, not the cloud copy

Zoom can transcribe for you, but its automated transcript is tied to cloud recording on paid plans — which means the entire meeting is uploaded to Zoom's servers and processed there. For a casual team sync that may be fine. For a sales call, a candidate interview, a client consult, an HR conversation, or anything under an NDA, sending the audio to a third party is exactly what you want to avoid.

Transcribing a local recording on your own machine sidesteps that. The Transcriber in Spark Tray runs entirely on your Windows PC — the audio is never uploaded — and it produces an accurate, timestamped transcript from the same recording Zoom already saved to your Documents folder. It's free, with no per-minute meter.

Step 1 — Record the meeting to your computer

Before the call, make sure you're recording locally, not to the cloud:

1. In Zoom, open Settings → Recording and confirm Local Recording is enabled.
2. Start the call, then click Record → Record on this Computer (as host, or with the host's permission).
3. Always tell participants you're recording — in many places one-party or all-party consent is a legal requirement, so get agreement up front.
4. End the meeting. Zoom converts the recording and saves two files: an MP4 video and an M4A audio-only file.

The M4A audio file is smaller and transcribes just as well as the video, so it's the convenient one to use.

Step 2 — Transcribe the recording privately

1. Open the Transcriber in Spark Tray.
2. Drag in the Zoom recording — the M4A audio or the MP4 video both work, with no conversion needed.
3. Pick the output format: SRT or VTT for time-coded captions, TXT for clean readable notes, or JSON if you want structured data with timestamps.
4. Run it. The transcript is generated on your machine; nothing is sent anywhere.

Accuracy comes from OpenAI's open Whisper speech-recognition model, which Spark Tray runs locally — so you get research-grade transcription without handing your call to a web service. Clear audio and minimal cross-talk help most; ask people to unmute one at a time when it matters.

Step 3 — Add speaker notes and labels

A local transcript captures words with timestamps but usually treats the audio as one stream, so speaker attribution is a manual pass — and a quick one when you have timestamps to navigate by:

• Keep the recording open beside the transcript and use the timestamps to jump to each turn.
• Add a name label at the start of each person's speech (e.g. Priya: / Marcus:).
• Clean up filler words and false starts only where readability matters — verbatim is better for legal or research records.

If your meetings are mostly one voice — a recorded lecture, a solo update, a dictated summary — attribution is a non-issue and the raw transcript is ready as is. The same workflow applies to any single-speaker audio, like transcribing voice memos.

Step 4 — Export and reuse

Once the transcript is right, export it for where it's going:

• TXT for minutes, notes, and pasting into email or a doc.
• SRT / VTT if you'll publish the recording with captions.
• JSON if another tool will ingest the timestamps.

Because everything ran locally, you can safely archive both the recording and the transcript wherever your data-handling rules require, without a copy sitting on a vendor's cloud. For the broader how-to on turning any recording into text, see how to transcribe audio to text.

The bottom line

To transcribe Zoom meetings well: record the call locally, then transcribe that file with an on-device tool so a sensitive conversation never gets uploaded. You get accurate, timestamped SRT/VTT/TXT/JSON output, easy speaker labeling, batch support for a backlog of calls, and full privacy — for free. Tell participants you're recording, keep verbatim where it counts, and your meetings become a searchable, quotable record you actually own.