How to add captions to a video on your computer

Captions help people follow a recording when they cannot hear it clearly or choose to watch without sound. They are especially useful in a software tutorial, where a short sentence can explain why you clicked a control instead of merely repeating its label.

Kliper Studio lets you create captions or import an SRT file, adjust the wording and timing, and save the result locally. Before sharing, it helps to understand the difference between captions drawn into a video, a separate caption file, and the editable project that produced them.

Trim and arrange your clips first

Remove obvious pauses and put your clips in the intended order before doing detailed caption work. Captions use the edited timeline, so a caption at 20 seconds belongs to that point in the project rather than automatically finding the same sentence in every source file.

Kliper keeps timed layers aligned when you keep or remove a range. Even so, review captions around each cut. A sentence may now begin in a different place, or a caption may remain too short to read comfortably.

If you are still combining recordings, follow the clip editing guide first. If the narration is missing, use the sound checklist; captions cannot restore audio that was never recorded.

Create a caption or import SRT

Open Captions in Studio. Move the playhead to the start of the sentence and choose Caption to create one. Enter the words and set the start and end times so they match the relevant part of the recording.

If you already have timed captions, choose Import SRT. An SRT file contains numbered pieces of text with time ranges. The current import accepts up to 100 cues. Confirm the file was made for the same version of your edit: a caption file for the uncut recording may not match a shorter timeline.

Play the first caption and one near the end after importing. That quick check can reveal a constant offset, a mismatched edit, or a timing problem before you spend time correcting individual words.

generate a first draft locally

In Studio, open Captions → Generate captions. This optional tool uses whisper.cpp and the multilingual Tiny speech model to make a first draft from the audio in your edited timeline.

Set up the tools once. On Windows and Linux, setup downloads a verified CPU caption engine and model, about 82–89 MB in total, depending on the platform. On Mac, choose an installed whisper.cpp whisper-cli executable first, then download the model, about 78 MB. Initial setup needs an internet connection. Keep extra disk space available for temporary audio while a job runs.

Choose a language and generate. Use automatic language detection or select the spoken language explicitly, then choose Generate captions. Recognition uses the edited clip order, trims, playback speed, and enabled audio-track levels. A track you have turned down or muted affects the audio being transcribed, so check the mix before you begin.

After the tools download, speech recognition runs on your computer; the recording does not need to be uploaded for this step. You can cancel a running job. Completion time and accuracy vary with the recording and computer, so try a short section before processing a long tutorial.

Review the transcript. Listen alongside the generated words and check names, numbers, punctuation, and timing. Quiet speech, unfamiliar terms, and overlapping voices can produce mistakes. Automatic captions are a first draft, and they still need your review.

Save the transcript or add it to the edit. Choose Save SRT file to keep the complete timed transcript, or Add captions to timeline to create editable cues. You can add up to 100 cues, subject to the project's shared limit of 120 visual layers. If the transcript is longer or the project has too many other layers, you can still save the full SRT file.

After adding captions, edit their text, timing, and appearance with the same caption controls described below. If you prefer not to install caption tools, manual captions and Import SRT remain separate options.

Correct the words, then the timing

Use the caption search to find product names, repeated terms, or a phrase you remember. Select a cue to edit it. You can split a cue at the playhead or merge it with the next cue when that makes the sentence easier to follow.

Write what the speaker means accurately. Include meaningful sounds or speaker identification when the viewer needs them to understand the recording. For example, “[error sound]” may matter if the narrator reacts to an alert that is otherwise invisible. W3C explains the role of speech and relevant non-speech information in its captions guidance.

Watch at normal speed while adjusting boundaries. Let viewers finish reading before the caption changes, and avoid covering the button or result the narration is discussing. When a line feels crowded, split the sentence rather than shrinking the type until it is difficult to read.

Make captions readable on the actual output

Choose a caption style, then check the picture behind it. A high-contrast background can help when the recording switches between light and dark windows. Keep a consistent size and position unless the screen content requires a change.

double-click an unlocked caption in the picture to edit its wording directly. Drag it to move it, and use the handles to resize or rotate it. Alignment guides help keep captions in a consistent position. A rotated caption may work as a label, but ordinary spoken captions are usually easier to follow when they stay level.

Preview the final canvas shape. A caption that looks small in a full-size editor can be hard to read after a landscape recording is viewed on a phone.

use text effects for a clear purpose

An entrance fade can soften a caption change. Typewriter reveals letters over a set duration. Word highlight changes the emphasized word as time passes.

Those effects are applied to the selected text or caption. Word highlight uses evenly spaced timing; it does not listen for each spoken word. Set the reveal duration, use the effect preview, and check it against the narration. If the animation distracts from a fast demonstration, use plain captions instead.

Save video, captions, and the editable project separately

Visible caption layers appear in the exported video picture. Viewers do not need to turn them on, but they cannot hide them either. Export captions .srt creates a separate timed-text file for a player or publishing service that supports it. Check that service's caption settings after upload.

Save a .kliper project when you want to change the wording or styling later. The MP4 is the finished picture; the project preserves the editable arrangement. If you move the project, copy its media and move the project file and assets folder together.

Check what leaves your computer

Local editing means you do not need to send the recording to a processing service for these caption tools. It does not mean every copy stays only on that device. A project saved in a synced folder may be uploaded by your sync software. Pro uploads send the selected video to your own YouTube or Google Drive account when you choose that action.

A project copied with media can include untrimmed source recordings and separate audio, not just the finished video. If you are sharing only the result, send the exported video and any caption file you intend to include. Review the privacy policy for the app and website's separate data handling, and check the actual files before sending them.