YouTube Subtitle Generator

Why it helps

Many viewers watch with the sound off, and others rely on captions to follow along. Uploading your own subtitle file means names and terms are right from the start, and a clean original track also gives YouTube's auto-translate better text to work from.

Step by step

1. Turn Full translation output off (text only) — you don't need a voice for subtitles.
2. Select the spoken language for untranslated subtitles, or another language for translated subtitles.
3. Download the SRT and fix names or terms in any text editor.
4. In YouTube Studio: Subtitles → Add language → Upload file → With timing.

Tips

• Run the file once per language you want; each run is 1 credit per minute (12¢ for a two-hour video).
• Use the timestamped transcript to write chapter markers for the description.

Example

Cost: 24 credits (2.4¢).

Limitations to know

• Subtitles don't label speakers or describe sounds.
• No URL import — use your own video file.

Plan one run per caption language

Each run of the video produces subtitles in one language, so the plan is simply a list of languages, each costing 1 credit per minute of video.

Captions in the spoken language

Select the spoken language itself; the text comes back untranslated. This is the track deaf and hard-of-hearing viewers use, and the text YouTube's own auto-translate works from.

Subtitles for another audience

Pick a target language. The SRT is machine-translated from the recognized speech and keeps the same timing.

A pair for language learners

Upload the spoken-language track and one translation so viewers studying your language can switch between them. Dual-language subtitles covers showing both at once.

Read the spoken-language file first. Translated runs start from the same recognition, so a word misheard there will be wrong in every translation as well. Note those spots and check them in each translated file before upload.

What an SRT cue looks like, and what to edit

An SRT file is plain text: a cue number, a timing line and one or two lines of text, followed by a blank line.

Open it in any plain-text editor and change only the text lines. The timing can stay as it is unless a caption appears visibly early or late. Things worth fixing before upload:

• Channel name, guest names and product names, which recognizers often spell by sound.
• Numbers and units, especially prices, measurements and model numbers.
• Lines that run long; subtitle readability suggests limits that keep reading comfortable.
• The file encoding: save as UTF-8, or accented letters may turn into garbled symbols.

The SRT editing guide has more on keeping cue numbering and timing intact.

Adding the parts recognition leaves out

The file contains the words that were spoken and nothing else. Captions written for deaf viewers usually also mark sounds that carry meaning, such as a door slam, a laugh after a punchline or music starting, and say who is talking when that isn't obvious from the picture. mydubly adds neither, so if your audience relies on captions, type short cues like [music] or a person's name before a line yourself. Captioning sound effects and music shows the common conventions. For a solo talking-head video you can usually skip this step; for a conversation or a comedy sketch, it decides whether the captions can be followed at all.

Chapters and descriptions from the same run

The same run also gives you a timestamped transcript, which is the quickest way to write chapter markers. Scroll through the timestamps, note where each topic begins, and copy those times with a short title into your description. YouTube reads chapter times from the description, so check its current rules on the first timestamp and the minimum number of chapters before publishing. The transcript also lets you lift an exact sentence for the description or a pinned comment without scrubbing through the video. Whether captions help a video get found is a separate question, covered in captions and YouTube SEO.

Uploaded translations versus auto-translated captions

YouTube can machine-translate a caption track for viewers on the fly, so it's fair to ask why you would upload translated files at all. The difference is review. An auto-translated caption is generated when someone watches, and you never see it before they do. A translated SRT is a file you can open, hand to a bilingual viewer, correct and only then publish. For a casual vlog the on-the-fly option may be enough; for a tutorial where a wrong measurement or a mistranslated button name sends viewers in the wrong direction, a checked file in your two or three biggest audience languages is worth the few credits.

Example: one tutorial, five caption tracks

If you want viewers to hear the video in another language rather than read it, the YouTube video translator page covers dubbing, and the SRT generator explains the file format itself.

Using the tool on this page

• The spoken language is detected automatically; you select the output language — the spoken language itself for an untranslated transcript, or another language for a translation. Files can be up to 2 hours long.
• Only the audio is sent for processing — the video stays on your device — and audio and text are deleted within 30 minutes of delivery (how files are protected).
• You can download a timestamped transcript, subtitles (SRT and VTT) and plain text. Every output, step and limit is explained on the video to text page.