How to add captions to a Twitter video people can follow

You recorded a useful explanation, but without the voiceover the clip is just a cursor moving around. Anyone who can't hear the audio has to guess what changed.

To add captions to a Twitter video, upload the video on the X/Twitter website, attach an .srt caption file beneath it, and review both before posting. X documents this as a web-only upload feature. Prepare the captions against the final edit so the words arrive with the action they explain.

Upload the caption file on the web

The steps below follow X's caption-file upload instructions:

1. Add your video to a post on the website. Beneath the video, select Upload caption file (.srt).
2. In the pop-up, select Upload captions and choose the caption file from your computer.
3. Select Done. Check that the caption language appears under the video in place of the upload command.
4. Review the post text, video, and attached captions, then publish when they're ready.

The uploaded captions appear near the bottom of the video when the viewer enables them. Attaching a file does not mean every viewer will have captions turned on.

If you're looking for this file-upload control in the mobile app, use the web flow documented above. If the website doesn't show the control described in X's help page, don't assume that typing a transcript into the post attaches a caption track. Check the current official instructions before publishing.

Choose a caption track or text baked into the video

An uploaded track lets the viewer show or hide the captions. Text baked into the video remains visible as part of the image. W3C calls these closed and open captions, respectively.

For an interview or spoken explanation, a caption track gives viewers that choice. If you want particular text to remain visible in the clip itself, add it during editing and inspect the exported video. Correcting baked-in text means editing and exporting the video again.

Avoid adding both versions by habit. Turn the caption track on during your check: two copies of the same sentence can compete with the thing you're demonstrating.

Captions convey speech and meaningful sounds, such as an alert needed to understand what happened. They don't automatically describe a visual action. A silent click on an unlabeled button still needs a clearer demonstration or an explanation. For a still image attached to a post, use image alt text rather than treating a video caption file as its description.

Build the file from the final audio

A transcript gives you words. A caption file also needs timing. Finish the cuts before adjusting captions; removing a pause from the video after export can leave the old file out of sync.

Use your editor's SRT export if it provides one, then inspect it alongside the exported video. This fictional example shows two numbered cues, each with a start time, an end time, and the words spoken in that interval:

1
00:00:00,000 --> 00:00:02,500
Open the saved version.

2
00:00:02,500 --> 00:00:05,000
The original file hasn't changed.

Those timings belong only to the example. Match your own cues to your recording. A plain paragraph saved with an .srt extension is not the same as a timed file.

Automatic transcription can give you a first draft, but listen through every line. Check product names and numbers carefully. Pay particular attention to short words that change the instruction: "has changed" and "hasn't changed" tell the viewer opposite things.

Break a long sentence where the thought naturally pauses. Read it at playback speed instead of judging it while the video is paused. If the text disappears before you can finish it, adjust the timing or the edit; shrinking the text won't give the reader more time.

Check the version a reader will watch

Open the final video at phone size and watch once with the sound off. Follow the task using only the visible information. When the explanation says "this option," you should be able to identify the option without replaying the moment several times.

Then check with audio on. Confirm that captions enter with the right phrase, remain readable, and clear before they obscure the next action. Inspect the bottom of the frame especially closely. A caption that covers the button being clicked makes the demonstration harder to follow.

After publishing, open the actual post and verify caption playback with captions enabled. Keep the video and its matching SRT together so you can identify which file went with which edit.

Caption accuracy is something you can inspect directly. A higher view count cannot establish it. The Twitter video views guide explains why post impressions, video playback, and understanding are separate questions.

Turn the response into your next explanation

Read the questions people ask after the clip. Repeated confusion about one step is a reason to explain that step more clearly, even if the post's total looks healthy.

PilotMyX keeps collected post text with available metrics. Review original posts from the last 30 complete days, inspect their wording beside impressions and available replies or bookmarks, and choose a comparable topic worth expanding. Keep missing metrics unknown and replies separate from original posts. A result on a different topic doesn't tell you whether captions helped this video.

Your connected agent can use that evidence and your description of the confusing step to prepare a text follow-up for review. Caption creation and playback checks stay in your video workflow. Start with the sentence your viewer needed to understand, then draft the next post around that explanation.