Manuscript Matching

Improve subtitle accuracy by providing a reference manuscript or script. The LLM uses this context to correct terminology, proper nouns, and domain-specific language.

What Is Manuscript Matching?

When you provide a text manuscript (script, transcript, or notes), VideoCaptioner uses it as a reference during subtitle optimization. The LLM cross-references the ASR output with your manuscript to:

  • Correct misheard proper nouns and technical terms
  • Maintain consistent terminology throughout the video
  • Fix domain-specific jargon that ASR often gets wrong
  • Preserve the intended meaning in ambiguous passages

How to Use

GUI

  1. In the subtitle optimization panel, find the Manuscript Hints section
  2. Paste your manuscript text, or load from a file
  3. Run the optimization — the LLM will reference your manuscript

CLI

Terminal
videocaptioner subtitle input.srt --manuscript script.txt --optimizer llm

Best Practices

  • Your manuscript doesn't need to be word-perfect — even rough notes help
  • Focus on including proper nouns, brand names, and technical terms
  • For presentations, export the slide text as your manuscript
  • For interviews, provide the interviewee's name and topic keywords

When to use manuscript matching

Most useful for technical content, lectures with domain terminology, or videos with unusual proper nouns. For casual content, the LLM usually does fine without a manuscript.