Auto-generated captions are usually 90% accurate and 100% unformatted: no punctuation, wrong keyword spellings, lines that break mid-phrase. That matters more than it seems — captions are indexed for search, read by a large share of mobile viewers, and required for accessibility. The Subtitle Formatter cleans and structures a caption file, enriches key moments with the correct keyword spellings, scores readability, checks accessibility compliance, and exports a translation-ready version.
This guide covers why captions affect discovery, the formatting rules that make them readable, and the translation path.
Captions and discovery
YouTube indexes uploaded caption text. If your spoken keyword is transcribed wrongly ("Gaggia" as "gaga"), the video loses that search signal. The formatter finds the topic terms you specify, corrects them throughout, and suggests where a spoken sentence could carry a keyword more clearly on the next recording.
Readability rules
Two lines maximum per caption, 32–42 characters per line, 1–6 seconds on screen, breaks at phrase boundaries not mid-clause, sentence punctuation restored, filler words removed. The readability score measures against these; the compliance check covers the accessibility conventions (speaker labels, sound descriptions) that matter for deaf and hard-of-hearing viewers.
Translation
A clean source caption file is the prerequisite for good translations, whether by a human or by YouTube's auto-translate. The translation-ready export strips timing quirks and non-speech artefacts so translated subtitles line up. Pair it with the Language Localization Finder to decide which languages are worth it.
Tips
- List your topic's proper nouns and technical terms in the input; those are what auto-captions get wrong.
- Upload the corrected file in Studio → Subtitles and set it as the default; it replaces the auto track.
- Re-use the cleaned transcript as the base for the description summary and a blog post.
Common mistakes
- Leaving auto-captions as-is — Wrong keywords and no punctuation cost search and retention.
- Long caption lines — Three-line captions cover the frame and are skipped.
- Skipping speaker labels in interviews — Accessibility fails and viewers lose the thread.
Worked example
Input: "Espresso tutorial; terms: Gaggia Classic Pro, Timemore, dial in, portafilter."
What the report returned: Corrections applied to 23 lines; readability 41 → 88; compliance passes; four keyword-enrichment suggestions for the next recording; SRT export ready for translation.
Frequently asked questions
Do captions affect ranking?
Caption text is indexed and improves topic understanding; accurate captions are a modest but real discovery signal.
Can I paste my SRT file?
Yes — paste the caption text or describe the topic and terms; the formatter returns a corrected, formatted version.
What about auto-translate?
It works far better from a clean, punctuated source. Format first, then translate.
Does this replace a professional captioner?
For most creator content, yes. For legal or medical content, use it as a first pass.