You finished the Webex call, the recording is sitting in your Webex site, and now somebody wants a transcript by end of day. Webex has a built-in transcription feature, but it only kicks in for certain plans, certain recording types, and certain languages. If yours didn't get one, or the one you got is unreadable, this is the workflow that actually gets you to a usable file.
The short version: download the MP4 from your Webex recordings page, run it through a transcription tool that handles speaker separation, then clean up the names. The whole thing takes about ten minutes for a one-hour call.
Does Webex transcribe meetings automatically?
Sometimes. Webex generates a transcript when the host enables it before the call and the meeting is recorded to the cloud (not locally). It's available on paid plans, with the cleanest support for English, Spanish, French, German, and a handful of others. If you recorded locally to your laptop, or the host forgot to toggle it on, there's no transcript waiting for you. You'll need to make one yourself from the MP4.
The built-in transcript also tends to miss speaker changes on calls with more than three or four people, and it doesn't always handle accents well. For interviews, depositions, or anything you're quoting from, most people end up redoing it anyway.
How do I download a Webex recording?
Sign in at your Webex site (usually yourcompany.webex.com), go to Recordings in the left nav, find the meeting, and hit the three-dot menu. Pick Download. You'll get an MP4 for video recordings or an MP3/M4A for audio-only.
If you're the host, you'll see every recording. If you were a participant, you only see what the host shared with you. No share link, no download — ask the host to forward you the recording or grant access.
For very long calls (anything over about 90 minutes), the download can stall on hotel Wi-Fi or VPN. Use a wired connection or split the download across a coffee break. The file is what it is — Webex doesn't offer a lower-bitrate version.
What's the fastest way to transcribe the file?
Upload the MP4 or MP3 directly to a transcription service. Most modern tools accept video files and pull the audio track automatically, so you don't have to convert anything. With VTS you can transcribe a video by dropping the file in, picking the language, and waiting. A one-hour call usually finishes in two to four minutes.
If you want to stay closer to the metal, you can run faster-whisper or OpenAI's Whisper locally. That's the right call if the meeting is confidential and you don't want the audio leaving your machine. The tradeoff is setup time and GPU memory — a one-hour call on CPU with the large model takes longer than the meeting did.
How do I get speaker labels for a multi-person Webex call?
This is where the built-in Webex transcript falls down. You need a tool that runs speaker diarization — the model that figures out when the voice changes and tags each chunk as Speaker 1, Speaker 2, and so on. From there, you do a quick pass at the top of the file to rename them: Speaker 1 becomes Priya, Speaker 2 becomes Marcus, and the find-and-replace handles the rest.
A few things that make diarization more accurate:
- Ask everyone to use a headset or a decent USB mic instead of laptop speakers. Cross-talk through built-in mics is the number-one reason speakers get confused.
- Mute when not speaking. This sounds basic. It still matters.
- If you can, record each speaker on a separate track. Webex doesn't do this natively, but if the call also went out on a podcast platform or Riverside, use those tracks instead.
The same playbook works for any video platform — we wrote a separate piece on the Zoom version of this problem with multiple speakers and most of the advice transfers cleanly.
How accurate is automated Webex transcription?
For a clean recording in clear American English with one or two people, expect somewhere in the high 90s on word accuracy. That's not 100. Names of people, products, and places are where the errors cluster, because the model has never heard them before. Acronyms get spelled out phonetically. Numbers spoken quickly get mangled.
For a five-person call with mixed accents, a HVAC system humming in the background, and somebody on a phone speaker, you should plan to spend 10 to 20 minutes cleaning up a one-hour transcript. That's still much faster than typing the whole thing yourself, which runs four to six hours of editor time per hour of audio at professional pace, according to the American Translators Association's guidance on transcription.
We wrote up what to expect from transcription accuracy in more depth if you're trying to set realistic expectations with a client or a boss.
Can I get timestamps for quoting from the call?
Yes, and you should ask for them up front. Most tools output two flavors: an SRT/VTT file with timestamps on every caption line (good for video editing), and a plain transcript with timestamps every paragraph or every speaker change (good for reading and citing).
For research interviews and legal work, the per-paragraph format is usually what you want — granular enough to find the moment, not so granular that the page becomes a wall of numbers. For captions you'd actually overlay on the video, you want the SRT.
What about confidentiality?
Webex meetings often contain things people don't want on a third-party server: HR conversations, customer data, vendor pricing, anything under NDA. Before you upload anywhere, read the transcription service's data policy. Look specifically for:
- Whether your audio is used to train models (it shouldn't be without opt-in).
- How long the audio and transcript are retained.
- Whether the data leaves the country it was uploaded in.
- Whether there's a Business Associate Agreement available if you're dealing with PHI.
If the answers don't satisfy your security team, that's when running Whisper locally makes sense, even with the extra friction.
Paste any public link or upload a file and get a clean transcript in minutes. First 3 clips every month are on us — no card required.
The verdict
For most Webex recordings, the fastest path is: download the MP4, drop it into a transcription tool that does diarization, rename the speakers, and ship the file. The built-in Webex transcript is fine for a quick recap of a small call and not much else. If the meeting is sensitive, run Whisper on your own hardware and pay the time tax.
Whichever path you pick, the trick is to ask for the right output format at the start. Going back to re-export with timestamps after you've already cleaned up the speaker names is the kind of small annoyance that eats an afternoon.
