Add Subtitles to a Twitter (X) Video

A timeline that plays on mute

On X, video normally starts playing by itself as it scrolls into view, with the sound off until someone taps. People can limit or turn off autoplay in their settings, but either way the clip begins as a silent picture in a fast-moving feed. Many viewers will decide whether to stop on your post before hearing a word.

That makes subtitles less of an accessibility extra and more of the message itself. A clip of someone talking, with no text, looks like a moving photo of a face. The same clip with a readable line on screen tells the scroller what's being said and gives them a reason to tap for sound, or to keep watching without it.

The caption options X gives you

X has offered three ways for words to appear on a video, and they behave differently.

• A caption file attached at upload. When posting from a web browser on a computer, X's media editor has offered an option to add an SRT file to a video. Viewers see those captions under X's control, and some apps show them automatically while the video is muted. The option hasn't been available in every app, and X has moved features around often, so look for it in the composer before relying on it.
• Automatic captions. X has generated captions for some videos in some languages. Coverage and accuracy vary, and you have little say over how they look.
• Subtitles burned into the video. Part of the picture before upload, so they show wherever the clip plays.

The file X has asked for is SRT. The editor's Export SRT writes standard SubRip text and timing, and if a hand-made file gets rejected, checking an SRT file for errors is the place to start. For background on the format itself, see what an SRT file is.

An SRT on X versus subtitles in the picture

Who sees them

SRT: viewers whose app shows captions, or who switch them on. Burned in: every viewer, every time.

Reposts, quotes and embeds

SRT: tied to that post on X. Burned in: travel with the video, including when someone downloads it or screen-records it and posts it elsewhere.

Look

SRT: X's caption style, varying by app. Burned in: your font, colour, box and position.

Changing a word later

SRT: depends on what X lets you edit, which has changed over time. Burned in: render again and post again.

Effort

SRT: export and attach, if your composer offers it. Burned in: one render on your computer before posting.

Neither column wins outright. An account that posts from the web and wants viewers to control text may prefer the file; a clip that's meant to travel, or that you'll also post to other networks, usually goes better with burned subtitles. What doesn't work well is both at once on the same post: a viewer whose app shows the SRT will see every sentence twice. The trade-off in general terms is in open versus closed captions.

Shapes that work on X

X accepts landscape, square and vertical video. How much of the frame shows in the timeline depends on the app and the screen, and a very tall clip may be shown in a shorter frame until someone opens it, so the top and bottom of a 9:16 video are the least safe places for text. Square and 4:5 clips fill more of a phone's timeline than 16:9 while still sitting comfortably on a desktop screen; subtitles for 4:5 and square video goes into those shapes.

The editor never changes the shape. Rendering keeps the source's resolution, frame rate and orientation, so export from your video editor at the shape you want on X first. The safe-area guide adapts to the frame: about 5% side and 7% bottom margins on 16:9, 6% and 9% on square, 6% and 10% on 4:5, and 8% and 16% on 9:16. It appears in the preview only; the rendered video carries just your subtitles.

X places a progress bar and a mute toggle near the bottom of the player, and the full-screen view adds its own controls. Upload limits on length and file size depend on the account type and have changed more than once, so check X's help for your account rather than trusting an old number.

Captioning a clip for X

1. Open the editor on this page and choose the clip from your computer: MP4, MOV, M4V, WebM or MKV. Style starts on Classic, white Inter text with a black outline, anchored to the bottom.
2. Get the words. Leave "Write the subtitles automatically with AI" switched on, or switch it off and use Import SRT / VTT, or press Enter at the playhead to type a line.
3. Correct names, handles and figures in the line list. Spoken @handles and product names are where speech recognition slips most.
4. Pick a style for the shape. Classic or YouTube suits 16:9; for vertical clips, switch to Social Media, which anchors to Middle and sits slightly below centre. Tick "Show safe area" and use "Nudge up / down" if a line meets a face.
5. If you plan to use X's caption upload instead of burning, click Export SRT and stop here.
6. Otherwise click Render video, confirm the summary, keep the tab open, then Download video and attach the MP4 to your post.

Writing for a two-second glance

Text on a timeline video competes with everything else on the screen, so a few habits matter more here than on a long video:

• Get a caption onto the first frame. AI lines begin on the first spoken word; if the clip opens on silence or a visual, place the playhead at 0:00, press Enter and type a short title line that says what's coming.
• Let the post text and the subtitles do different jobs. The post frames the clip; the subtitles carry what is said. Repeating one in the other wastes the glance.
• Keep lines short. AI subtitles come out as single lines of about 32 characters, and the editor warns when a line needs more than 21 characters per second to read.
• Write numbers as digits and check every proper noun, because a misheard company name in a public post gets screenshotted.
• Add sound cues by hand when the clip depends on them. The AI doesn't write [laughter], [music] or speaker names; the SDH subtitles page shows how to format them.

One last check, then post

Play the downloaded MP4 on a phone with the sound off and judge it as a stranger scrolling past would. If the first line makes sense on its own and nothing hides under the player's controls, it's ready. For an audience that reads another language, set "Subtitle language" before running AI subtitles, or use the video translator for a full translated version. If all you want is the subtitle file, the subtitle generator skips the render, and burning subtitles into a video explains the render step in more depth.

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