Removing Captions from Zoom, Teams, and Google Meet Recordings Before You Reuse Them

A good explanation from last month's webinar can become a two-minute tutorial. A clean answer from a class replay can become a promo clip. Then you open the file and find a caption band across the speaker's chest, a name tag in the corner of every tile, and a date stamp nobody asked for.

Some of that text is not in the video at all. Zoom, Microsoft Teams, and Google Meet store captions and transcripts as separate data in most of their own recording workflows, so the fix is a toggle, not an edit. Other text, such as Zoom's optional name labels and timestamps, or anything captured by a screen recorder pointed at the meeting window, is part of the picture. Those pixels have to be removed, covered, or cropped.

This guide sorts the common text layers by platform, shows how to check your own file, and then walks through cleaning the burned-in ones with RemoveText.video, including where that cleanup is likely to struggle.

Quick answer: which meeting text is actually in the video?

Start here. The table summarizes what each platform's official documentation says as of November 2026. Settings move, so check your own account before relying on it.

One row is missing from every vendor's table: a third-party screen capture. If someone recorded the meeting with OBS, QuickTime, the Xbox Game Bar, or a phone pointed at a laptop, the capture contains whatever was on their screen. Live captions, name tags, "you are sharing" bars, reaction emoji, and chat previews are all baked into the frames.

The text layers, one at a time

Live captions and transcripts

Captions are the most common complaint and the one most often solved without editing.

Zoom. In the web portal's cloud recording settings, "Save closed caption as a VTT file" stores captions as a VTT file on the recording's detail page. A separate setting, "Create audio transcript," produces an audio transcript. Zoom's transcription article says the transcript "is automatically embedded within the audio and video file, but is hidden by default," and appears on top of the video only when a viewer clicks the CC icon. Zoom also notes that audio transcription applies to cloud recordings, not computer recordings.

Teams. Microsoft is direct about this one: live captions "aren't captured in recordings or downloads." Captions you see during playback in Teams or the Microsoft 365 video player come from the meeting transcript, or from captions generated later on the video's player page, and viewers switch them with Toggle captions (Alt+C). If a downloaded Teams recording shows caption text inside the picture, something other than Teams' recorder put it there, most likely a screen capture or a re-export with burned-in subtitles.

Google Meet. When you start a recording from Meeting tools > Recording, you can choose a caption language and record captions. Google's help says this will "embed captions in the clip," but its playback instructions point to the Closed Caption button in Drive's player, or to enabling subtitles in VLC or QuickTime after download. That describes a caption track that can be switched off, not lettering in the image. Confirm it on your own download with the check below. Google also warns the captions may become playable a few hours after the video file.

A quick test settles it. If the words vanish when you turn subtitles off in a desktop player, you have a track, not pixels. Drop the track when you export your edit and keep the text file for later.

A presenter faces her laptop webcam; the screen shows her video tile next to a shared mountain photograph.

Participant name tags

Zoom's cloud recording setting "Display participants' names in the recording" adds each participant's name to the bottom-right corner of their video. When it is on, the names are part of the picture. It is a recording setting, so clearing it changes recordings made afterward, not files that already exist.

Neither Microsoft's nor Google's recording articles say whether name labels appear inside the recorded video. Don't guess. Scrub through your own file at a moment when several people are visible.

Name tags matter for more than looks. A tutorial or promo clip may reach people who never agreed to have their name published next to their face. Removing a label does not anonymize anyone, though. The face and voice are still there, so get consent for the reuse itself.

Timestamps

Zoom offers "Add a timestamp to the recording" for cloud recordings, with the time shown in the host's time zone. University Zoom guides document a matching option for computer recordings in the desktop app's Recording settings. Either way, the stamp is drawn into the frames. Turn it off in both places before the next session.

Screen-share notices, presenter thumbnails, and other pop-ups

Here the recording method matters more than the platform.

Zoom says its computer (local) recordings "reflect the view of the person recording," including pins, spotlights, and their current layout. Cloud recordings follow the host's settings instead. For reuse work, the most useful Zoom option is the separate shared-screen file. Turn on "Record active speaker, gallery view and shared screen separately," then select Shared screen, and you get the slides without anyone's face on top. Clearing "Record thumbnails when sharing" (Zoom's layout article calls it "Record tiles when sharing") keeps the presenter thumbnail off the shared content. All of this has to be set before the meeting starts.

Google says Meet recordings include the active speaker and anything presented, and that other windows and notifications aren't saved. Microsoft lists what Teams recordings do not capture, such as whiteboards and annotations, shared notes, and more than four video streams at once.

A screen recorder ignores all of that. Its output includes the meeting app's own controls, any "You are screen sharing" bar, and whatever notification slid in during the call.

Text inside the shared slides

Slide titles, bullet points, and code on a shared screen are content, not overlays. If they are wrong or confidential, fix the deck and re-record the segment, or cut it. Rebuilding letters with a cleanup tool can produce something that looks like text but isn't accurate.

Check your file before you edit anything

Two minutes of checking can save a processing job.

1. Play the downloaded file in VLC or another desktop player. Open the subtitle menu and disable every track. Anything still visible is in the pixels.
2. List the streams. If you have FFmpeg installed, ffprobe shows whether the file carries a subtitle stream alongside video and audio:

ffprobe -v error -show_entries stream=index,codec_type,codec_name -of compact meeting.mp4 

A line with codec_type=subtitle means a switchable track exists. No such line, yet text on screen, means burned-in text.

1. Look for an alternate file. Zoom cloud recordings can come with separate active-speaker, gallery, and shared-screen files. One of them may already be clean.
2. Note every layout change. Write down when the view switches between speaker, gallery, and shared screen, and when people join or leave. Name tags and caption bands move with those changes, and this list becomes your selection plan.

If the text is a track, stop here. Export without it, or hide it in the hosting player.

How to remove burned-in captions and name tags with RemoveText.video

For text that really is in the frames, use a tool that rebuilds the background under a selected area instead of cropping it away. Use footage you own or have permission to edit, and work on a copy.

1. Cut a short excerpt first

Meeting recordings are long, and processing is billed by the second. Trim the clip you actually plan to reuse in your editor, and leave a second or two on either side of each cut. RemoveText.video accepts MP4, MOV, and WebM files up to 100 MB, with a short side from 240p to 2160p.

The cost is easy to estimate before you commit. The rates listed on the caption removal page are 1 credit per second for a 240p–720p short side and 2 credits per second for 721p–1080p. A 30-second excerpt from a 1080p recording is about 60 credits. New accounts get 20 one-time trial credits after Google sign-in, enough for roughly 10 seconds at 1080p, which is plenty for a test on the hardest moment.

2. Open the clip for a local preview

The upload box reads "Drop a video here," lists MP4, MOV, or WebM up to 100 MB, and notes "Preview locally before upload."

Drop the excerpt onto the RemoveText.video homepage. Selecting a file creates a preview in your browser. Nothing uploads until you confirm processing. You can mark areas without signing in, and sign-in is needed only when you process.

3. Choose Text and draw a tight box around each layer

The editor shows Area 1 drawn over on-screen text, a Target choice of Text or Watermark, an Areas to remove counter, a Trim video panel, and an estimated cost.

Under Removal settings, pick Text as the target. The panel describes it as the option for subtitles, titles, and captions. Then drag on the video to draw an area. Each area appears in the Areas to remove list, and you can add more by dragging again. Free accounts can mark up to three areas, and paid plans up to ten.

In a meeting recording, that usually means:

• One box for the caption band, sized to the widest line that appears in the excerpt, not the line on the current frame.
• One small box per name tag that needs to go.
• A box for the timestamp, if Zoom added one.

Keep each box close to the text. The selected area is the only place cleanup happens, and RemoveText.video does not look for extra targets outside it. A larger area also means more background to rebuild and a greater risk of touching a face or a slide. If a caption sits on a dark backing panel, include the whole panel. Removing only the letters can leave the panel behind.

4. Review the estimate and process

Check the estimated cost and processing time at the bottom of the panel, then choose Remove text. The note under the button reads "Your video uploads only when you start processing." If your balance does not cover the whole clip, the workbench shows how much it can cover, so you can process a shorter range first and compare the result before buying more credits.

5. Watch the entire result

Before-and-after sliders on the RemoveText.video homepage: product text removed from a studio shot, and a burned-in caption removed across a speaker's clothing.

Play the output from start to finish at full size. Then scrub slowly through every layout change you noted earlier. Look for smearing on clothing, shimmer where the caption used to be, and edges that pulse when the speaker moves.

RemoveText.video's processing standards say the output is re-encoded. Timing, frame-rate representation, codecs, and audio tracks can differ from the source, and duration can differ slightly. Drop the result back on your timeline and check sync at the start and end before cutting it into the final edit. Originals are kept for up to 7 days, and outputs for 7 days on the trial or 30 days on paid plans, so download what you need.

Where cleanup struggles in meeting footage

Meeting recordings are harder than they look. Webcam footage is often dim and compressed, and in speaker view the frame changes whenever someone else starts talking. Expect trouble in these cases.

A paused close-up of a speaker raising his hand near his chin, shown on a monitor above a keyboard and a jog controller.

Captions over faces, hands, and microphones. In speaker view the caption band often crosses a chin, a gesturing hand, or a headset boom. RemoveText.video's own documentation names fast motion, faces, hands, hair, and fine texture as areas that can leave flicker, blur, or reconstruction artifacts. The tool's caption page also says a clean source is the better choice when old text covers a face or hand.

Captions over busy slides. If the caption sat on top of a shared slide full of text, the tool has to invent what was under it. The result may look plausible without being right. For teaching content, re-export the slide from the deck and rebuild that section, or use Zoom's separate shared-screen file if you have it.

Moving tiles and speaker switches. A selection box does not follow a tile. In gallery view, a name tag moves when someone joins, leaves, or turns their camera off. In speaker view, the frame cuts to a different person mid-sentence. Split the excerpt at those points and process each segment with its own boxes, or pick excerpts that hold one layout.

Opaque caption boxes. A solid box hides everything beneath it, so there is nothing left to recover. On a plain wall that may still work. On a patterned shirt or bookshelf, the repair can smear. Cropping, or covering the area with a new lower third, may look better.

Low resolution. Google notes that a Meet recording's resolution depends on the largest shared screen, up to 1080p. Webcam tiles inside a gallery grid are far smaller than that. A blurry, blocky source gives the cleanup very little to work with.

If a section still distracts after one or two tries, change the edit instead. Use a tighter crop, a cutaway to the slides, a new lower third placed over the old one, or a short re-recorded segment.

Put accessible captions back on the finished clip

Removing captions from a recording should not leave the published clip without them. The W3C's captions guidance says captions should include relevant non-speech audio as well as dialogue, and that automatic captions need checking for accuracy. Zoom's VTT transcript is a decent starting draft, and Teams and Meet transcripts can serve the same purpose. Edit the text for the excerpt and fix speaker names. Then attach it as a caption file in your editor or on the hosting platform, so viewers can switch it on and off and the next person who reuses the clip doesn't run into the same problem.

Set up the next recording so there is less to remove

A few settings before the meeting do more than any cleanup afterward:

• Zoom (web portal, Recording & Transcript): clear "Add a timestamp to the recording" and "Display participants' names in the recording." Turn on "Save closed caption as a VTT file" and "Create audio transcript" so captions arrive as files. Turn on "Record active speaker, gallery view and shared screen separately" with Shared screen selected for a clean slide track.
• Zoom (desktop app, Settings > Recording): for computer recordings, clear the timestamp option there too.
• Teams: start transcription with the recording. Playback captions then come from the transcript, and Microsoft says captions aren't captured in recordings or downloads.
• Google Meet: use Record captions if you want captions you can switch off later. Present from the meeting itself rather than screen-recording the browser.
• Everyone: avoid third-party screen captures of the meeting window when the built-in recorder is available. If you must use one, turn captions off on that machine and close chat and notification panels first.

FAQ

Can I turn off Zoom names or timestamps in a recording that already exists?
No. Those settings apply when the recording is made. For an existing file, remove the area visually, crop it, or cover it.

Why does my Teams recording show captions in the browser but not in the downloaded MP4?
Because they were never in the video. Microsoft says captions aren't captured in recordings or downloads. The player displays them from the meeting transcript or from captions generated on the video's player page.

Is removing a name tag enough to make a clip anonymous?
No. The person's face, voice, and anything they say about themselves remain. Get permission for the reuse, or cut or blur the person.

The short version

Check before you edit. Captions in Zoom, Teams, and Meet recordings are usually a separate file or track. Zoom's name labels and timestamps, and anything a screen recorder captured, are pixels. For those, trim to the excerpt you need, draw tight boxes around each layer, watch the whole result through every layout change, and switch to a crop, a cover, or a re-record wherever the repair crosses a face or a slide.