agentleFS
Sign inStart free
AbdalkaderdevAbdalkaderdev/takecut/skills/video-captions/SKILL.md

video-captions

Skill0 starsChanged 35 days agoMITView on GitHub

Make accurate subtitles and styled burned-in captions from word-level timestamps. Generates SRT and WebVTT that follow broadcast rules (42 characters per line, 2 lines, reading speed, minimum duration, gaps), ASS captions with presets (clean, boxed, minimal, word-highlight karaoke, TikTok-style pop), burns them into video, and translates captions while keeping the original timings. Use when the user asks for subtitles, captions, SRT/VTT files, burned-in or animated captions, or subtitles in another language.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. kitcut-hq/kitcut★ 42same name
    2. Muvon/octomind-tap★ 6same name
    3. Sanidhya14321/video-forge★ 2same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository4 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Always report output paths and the cue count.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

video-captions

Script paths are relative to this skill's folder; run them by full path from the user's working directory (don't cd into the skill). Input is a words JSON: either edits/<name>/transcript.words.json (captions for the original) or the <edit>.words.json that video-edit's render.py writes next to an edited output (captions for the edit; no re-transcribing).

Files

  • SRT: uv run scripts/captions.py srt WORDS.json -> <stem>.srt
  • VTT: uv run scripts/captions.py vtt WORDS.json
  • Styled ASS: uv run scripts/captions.py ass WORDS.json --preset highlight --video VIDEO.mp4 (or --size 1080x1920; PlayRes matches the video so sizes are proportional)
  • Burn in: uv run scripts/captions.py burn VIDEO.mp4 SUBS.ass -> VIDEO.captioned.mp4 (an .srt is converted with the clean style). Audio is copied untouched.

Missing transcript: uv run ../video-edit/scripts/transcribe.py VIDEO first.

Presets

presetlookuse
cleanwhite bold, black outline, bottomdefault landscape subtitles
boxedwhite on a translucent black boxbusy backgrounds, accessibility
minimalsmaller, thin outlinedocumentary, interviews
highlightfull line, active word yellowtutorials, talking heads
pop1-3 uppercase words, active word green, slight scale pop, lower-middlevertical shorts

Options: --font "Inter" (must be installed, or pass --fonts-dir to burn), --max-chars, --max-lines, --max-cps (default 17), --min-dur (0.83 s), --max-dur (7 s). Line-break and timing rules: references/rules.md.

Translation (timings preserved)

  1. Make the source-language SRT (above).
  2. uv run scripts/captions.py translate-prep SUBS.srt -> SUBS.lines.json ({"1": "text", ...}).
  3. Translate every value yourself into a new file, e.g. SUBS.de.lines.json. Keep every key, one entry per cue, no merging or splitting. Translate meaning, keep it about as short as the original (subtitles are read fast); keep names and numbers.
  4. uv run scripts/captions.py translate-apply SUBS.srt SUBS.de.lines.json --lang de -> SUBS.de.srt with the original timings, re-wrapped to the line rules. It refuses if keys don't match. Use --lang zh|ja|th for languages written without spaces.
  5. Optional: uv run scripts/captions.py ass SUBS.de.srt --preset boxed --video VIDEO.mp4, then burn.

Always report output paths and the cue count.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free