Add Subtitles to a Screen Recording

Subtitles on top of other text

In most videos subtitles sit over a picture: a face, a street, a product. In a screen recording they sit over more words. Menus, code, form fields, chat panels and terminal output fill the frame, and the thing a viewer needs to read is often exactly where captions normally go. That changes the priorities. Position matters more than style, because the right spot differs from one recording to the next. The subtitles also compete with the interface for the viewer's eye, so they need to look clearly unlike it without being loud.

There is a third issue: screens are captured at high resolution and then watched in a small player or on a phone, so interface text may already be close to unreadable. Subtitles can't fix that; recording at a larger zoom level can. What subtitles can do is avoid covering it.

The files recorders produce

Recording tools save different formats, and the editor opens MP4, MOV, M4V, WebM and MKV. Defaults change between versions, so check your recorder's output setting rather than relying on this list:

OBS Studio

Commonly set to MKV or MP4. Either opens directly; the MKV page explains what happens to an MKV's own tracks.

macOS screenshot toolbar and QuickTime Player

MOV files.

Xbox Game Bar on Windows

MP4 files, usually in a Captures folder under Videos.

Browser extensions and web-based recorders

Usually WebM with VP8 or VP9 picture and Opus sound; see subtitles for WebM recordings for those specifics.

Meeting apps

Recordings normally arrive as MP4.

Many screen recorders write variable frame rate, saving new frames mainly when something on screen changes. The editor keeps every frame's original timestamp and places subtitles by time, so a recording that idles for ten seconds and then scrolls quickly stays in sync. The summary shows an average frame rate; variable frame rate explains the idea.

Choosing a spot that covers nothing important

Scrub through the recording first and notice where the action happens, then pick the anchor:

• Bottom, the default for most presets, when the important part is the upper or middle screen: slides, dashboards, documents, design canvases.
• Top, when the bottom carries the action: a terminal or console panel, a status bar, a chat box, the Dock or taskbar, or the timeline of a video or audio app. The Top anchor puts the first line just inside the top margin.
• "Nudge up / down" when the chosen band is almost clear but not quite. It shifts the whole block by a share of the frame height; positive values move it up.
• A smaller "Max line width" when the action is at one side. A narrower centred block covers less of a sidebar or a code gutter.

Position applies to the whole video rather than to single lines, so choose the spot that works for the stretches that matter most, and keep lines short where they would overlap something. Moving subtitles to the top goes further into anchors. "Show safe area" draws the margins for the frame's shape; on a 16:9 capture they are slim, so the text can sit close to the edges.

Readable over interface text

Minimal, the preset this page opens with, sets Inter at a medium weight, a little smaller than the default, with a soft dark shadow and no box. It reads well on dark themes, code editors and video players. On light interfaces such as documents, spreadsheets and most web pages, white text with only a shadow can blend into the page. Switch the background to Box, or pick the YouTube preset (white on a dark box) or High Contrast, and the lines read on anything. Subtitle backgrounds compares box, outline and shadow side by side.

Size is a share of the frame's shorter side, so it doesn't shrink on a big capture: on a 2560×1440 recording Minimal sets the font at about 66 pixels, the same share of the frame as on a 1080p capture. If people will watch on phones, raise the size rather than lowering it. It also helps when captions look nothing like the app being shown; a font and colour that differ from the interface's own text tell viewers at a glance which words are narration.

Captioning a tutorial from start to finish

1. Open the editor and choose the recording. Its resolution and frame rate appear above the preview once it has been read.
2. Bring in the narration: AI subtitles if you talked while recording, Import SRT / VTT if a script or earlier file exists, or type lines with Enter at the playhead.
3. Correct product names, menu labels, shortcuts and commands so they match the screen exactly. Copy them from the interface rather than spelling them from memory.
4. Open Style. Keep Minimal for dark screens, or switch the background to Box for light ones.
5. Choose Top or Bottom, then scrub to the busiest moments and use "Nudge up / down" until nothing you point at is covered.
6. Look through the line list for warnings. Long technical terms push lines past a comfortable reading speed quickly.
7. Click Render video, start it, then Download video. Export SRT too if the tutorial is going to a platform that accepts caption files.

Jargon, silence and noise

Speech recognition is weakest where tutorials are densest: product names, keyboard shortcuts, file paths, command-line flags, acronyms and version numbers. A function name may come back as ordinary words. The fix is usually quick, because the term you were saying is visible on screen; fixing auto-generated subtitles suggests an order for the review.

Tutorials also contain long silences while you click, type or wait for something to load. The AI writes lines only where there is speech, so those stretches stay free of text. If a silent step needs explaining, add a short caption there by hand, such as "Saving the settings".

Microphone choice decides most of the accuracy. A headset or desk microphone beats a laptop's built-in one, and keyboard clatter right beside the mic causes misheard words; screen recording audio settings covers the setup for next time.

Recorded meetings sometimes already show the meeting app's live captions in the picture. "Hide subtitles already in the video" blurs a horizontal band where those sit so your corrected lines can be drawn on top; everything inside the band is blurred, so place it carefully.

Three recordings, three layouts

For a subtitle file without rendering, the AI subtitle generator is the direct route, and translated subtitles produces one in another language. Teams publishing demos abroad will find more in localizing product demo videos, and recorded classes have their own page on lecture captions.

About the subtitle editor on this page

Opens

MP4, MOV, M4V, WebM, MKV videos your browser can decode

Subtitles from

AI (from the speech, optionally translated into one of 21 languages), an SRT or VTT file, or typing them in

You download

An MP4 with the subtitles drawn into the picture at the original resolution and frame rate, plus SRT and VTT files

Cost

Editing, styling, rendering and SRT/VTT export are free, with no account and no watermark. AI subtitles cost 1 credit per minute (5-credit minimum per video); signing in with Google gives 100 free credits and accounts get 10 free credits a day

Privacy

The video never leaves your device and is rendered in your browser. Only for AI subtitles is the audio sent, over HTTPS, and it is deleted within 30 minutes of delivery

Rendering needs

Chrome or Edge 94+, Safari 16.4+, or Firefox 130+ on a computer. Editing and SRT/VTT export work in any current browser