How to add captions to a YouTube video with an SRT file
YouTube can generate automatic captions, accept a timed subtitle file, synchronize a plain transcript, or let you type captions in its editor. A reviewed SRT file gives you control before publication and leaves you with a portable caption master that can also be used in an editor or another platform.

Choose the caption method before you begin
Best when you want to create and review both words and timing before opening YouTube Studio.
YouTube can auto-sync text in the same language as the spoken dialogue, but timing still needs review.
Convenient after upload, but names, jargon, punctuation, and sound descriptions require a deliberate pass.
This guide uses the timed SRT route because it keeps a reusable copy outside the platform. YouTube also accepts other caption formats, including basic WebVTT, but its official supported caption-file list recommends simple formats such as SubRip for creators getting started.
1. Work from the source video, not the watch page
Create the caption file from the same video master you upload to YouTube. A YouTube watch-page URL is a streaming webpage, not a direct media file that CaptionKite can import. If the video is already published, use an authorized local copy or a download provided through the account and platform controls available to you.
Do not create captions from a differently edited version. Removing a five-second introduction, changing the speed, or inserting a title card shifts every cue that follows. Lock the picture first, then make and upload captions. If the final video changes, regenerate or retime the track rather than hoping the difference is too small to notice.
Play the master once and note the spoken language, unusual names, intentional silences, music, and speakers who are not visible.
2. Create an SRT file in CaptionKite
- Open the CaptionKite caption tool and select the final local video.
- Choose the language actually spoken in the recording.
- Generate the caption draft and keep the tab open while it processes.
- Review the text and timings against the video.
- Download the corrected result as SRT.
CaptionKite transcribes a selected local file in the browser, so the media is not uploaded to CaptionKite. The downloaded SRT is plain UTF-8 text divided into numbered cues with start time, end time, and caption text. That simplicity is why it travels well between editors and publishing systems.
If SRT structure is new to you, follow how to create an SRT from video before continuing.
3. Review words, timing, and useful sounds
Start with names, organizations, products, locations, dates, measurements, prices, URLs, and acronyms. Automatic speech recognition often replaces an unfamiliar proper noun with an ordinary word that looks completely plausible. Then check punctuation and line breaks without rewriting the speaker's meaning.
- Keep captions on screen long enough to read.
- Break lines at natural phrase boundaries.
- Avoid one-word flashes during rapid speech.
- Identify an off-screen speaker when the viewer cannot infer who is talking.
- Add useful sounds such as [music], [applause], or [door closes] when sound carries meaning.
- Check the first and final cues, which are easy to miss during spot review.
Play the complete video with captions once before upload. A structurally valid SRT can still contain a late cue, a missing sentence, or text that covers an important visual moment.
4. Upload the SRT in YouTube Studio
Google's current YouTube caption instructions use this path:
- Sign in to YouTube Studio.
- Choose Subtitles from the left menu.
- Select the video.
- Add the spoken language if it is not already listed.
- Under Subtitles, choose Add.
- Select Upload file, choose the option with timing, and select the reviewed SRT.
- Inspect the imported cues, then save or publish them.
Interface labels can change, and captions may also be added during the video-upload flow. If the page differs from these steps, use YouTube's current help link above rather than searching for a third-party workaround. Confirm that you are editing the correct channel and video before publishing.
5. Label the language and track correctly
The language setting is not decorative. It helps YouTube and viewers understand what the track contains, supports language selection, and separates an original-language caption track from translated subtitles. Choose the language of the spoken dialogue for the source track.
If a recording contains occasional phrases in another language, keep one primary track and transcribe those phrases accurately where appropriate. If the entire video needs a translated track, create it as a separate subtitle version and have a fluent reviewer compare it with both the original speech and the intended audience's language.
Use a clear track name when Studio offers one, especially for distinctions such as English and English (Canada), edited captions, or descriptive captions. Do not label machine-translated text as reviewed merely because the file uploaded successfully.
Automatic captions are a draft, not a publishing decision
YouTube may publish automatic captions when supported. Google explicitly advises creators to review them because speech recognition can misrepresent accents, dialects, pronunciation, and noisy audio. Its current automatic-caption guidance also lists long videos, silence, poor sound, overlapping speakers, and mixed languages among reasons captions may be delayed or unavailable.
A custom reviewed track is valuable even when automatic captions already exist. It lets you correct names before viewers encounter them, include meaningful non-speech audio, preserve a copy outside YouTube, and use the same approved wording in other publishing channels.
If you edit an automatic track in Studio, verify that the corrected version is the active published track and that an outdated duplicate is not confusing viewers.
Understand what an SRT file does not control
Basic SRT carries text and timing. It does not reliably control fonts, color, animation, or exact screen position on YouTube. The player presents selectable captions according to platform behavior and viewer preferences. This is an accessibility advantage: viewers can turn captions on, change how they look, or use the language menu.
If the words must always appear in a specific style, they have to be burned into the video in an editor. Burned-in subtitles cannot be turned off and are not a substitute for a proper caption track where accessibility matters. Some creators publish both: carefully placed visible text for silent browsing plus an accurate selectable track for viewers who use caption controls.
Read selectable versus burned-in captions before committing to permanent text.
6. Check the published result as a viewer
After processing, open the public or unlisted watch page in an ordinary viewer session. Turn captions on with the CC control, select the intended language, and play the beginning, a fast section, and the ending. Check on a phone as well as a desktop because line wrapping and the amount of visible picture differ.
- Confirm the track is available rather than still a draft.
- Look for cues hidden by other on-screen text.
- Check that rapid cues do not flash too quickly.
- Verify music and sound descriptions are useful and not excessive.
- Make sure a translated or alternate track has the correct label.
If an error appears later, YouTube Studio allows text and timestamp editing. For extensive changes, download or update your reviewed master too; otherwise the platform copy and your archive drift apart.
Fix captions that drift or appear at the wrong time
If every caption is early or late by the same amount, the SRT may have been made from a version with a different introduction. If timing starts correctly and gradually drifts, the source and uploaded video may differ in frame rate, speed, or duration. Recreate the captions from the final master rather than moving hundreds of cues individually.
If the file is rejected, open it in a plain-text editor and confirm that it contains numbered cues, valid start and end times, and ordinary UTF-8 text. Do not rename a TXT transcript to .srt; the internal structure matters. Remove unsupported styling and test a small corrected file if necessary.
When only a few cues are wrong, Studio's editor can adjust their timestamps. Keep the local SRT in sync with those corrections if it is your reusable master.
Keep privacy and publishing rights separate
CaptionKite's local processing avoids uploading your source to a transcription server, but publishing on YouTube deliberately sends the video and caption track to YouTube. Review both for private names, addresses, account details, confidential screen content, and conversation that was not intended for a public audience.
Use only video you own or are authorized to publish. A caption file does not grant permission to reuse the speech, music, or images in a video. If other people are identifiable, follow the consent and organizational rules that apply to the recording.
The transcript can reveal information that is hard to notice while watching. Search the text for phone numbers, email addresses, client names, and private project terms before uploading.
YouTube caption publishing checklist
- Use the exact final video master.
- Select the correct spoken language in CaptionKite.
- Correct names, numbers, punctuation, and sound descriptions.
- Preview the complete SRT with the video.
- Upload it to the correct video and language in YouTube Studio.
- Choose the upload option that preserves the existing timing.
- Publish the track rather than leaving it as a draft.
- Open the watch page and test captions on desktop and mobile.
- Keep the reviewed local SRT and update it when platform edits are made.
- Repeat the review after replacing or re-editing the video.
For captions intended for several destinations, build one reviewed master and then follow the social-media subtitle guide for platform-specific layout decisions.