agentsclimarketplace

Daftai extracting video subtitles

Skill daftAI2026/daftAI-skills/skills/daftai-extracting-video-subtitles

Install
npx -y skills add daftAI2026/daftAI-skills --skill daftai-extracting-video-subtitles

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Extracts timed subtitles from local video/audio files using Whisper. Outputs SRT format with accurate timestamps. Handles complex filenames automatically. Use when user asks to "extract subtitles", "get subtitles", "transcribe video", "识别字幕", "提取字幕", or "获取字幕".

SKILL.md

3.0 KB, as published. Nobody here has run it

Video Subtitle Extractor

Extract timed subtitles (SRT) from local video/audio files using OpenAI Whisper.

Scripts: All scripts are in scripts/ relative to this SKILL.md.

Workflow

- [ ] Step 1: Environment check (Whisper + FFmpeg)
- [ ] Step 2: Input validation (file exists, has audio stream)
- [ ] Step 3: Detect language (auto or user-specified)
- [ ] Step 4: Run extraction script
- [ ] Step 5: Verify output and report

Step 1: Environment Check

which whisper
which ffprobe
ResultAction
whisper foundContinue
whisper missingbrew install openai-whisper (macOS) or pip install openai-whisper
ffprobe missingbrew install ffmpeg

Step 2: Input Validation

  1. Confirm video/audio file exists
  2. Get duration and confirm audio stream present — handled by script automatically

Step 3: Language Detection

  • If user specifies language → use it
  • If unclear → ask user, suggest en or zh as common options
  • Whisper --language flag values: en, zh, ja, ko, fr, de, es, etc.

Step 4: Run Extraction

bun scripts/extract_subtitles.ts "<video_path>" [language] [model]

Parameters:

ParameterDefaultDescription
video_path(required)Input video/audio file
languageenLanguage code
modelauto by durationWhisper model

Model auto-selection (by duration):

Video durationModelApprox. speed (CPU)
< 2 minmedium~2-3 min
2-30 minturbo~1x duration
> 30 minbasefastest

The script automatically handles:

  • Complex filenames (copies to temp dir)
  • Audio stream validation
  • Duration detection and model selection
  • Temp file cleanup

Known issues:

IssueSolution
FP16 not supported on CPU warningNormal on Mac CPU, auto falls back to FP32
MPS NaN errorsDo NOT use --device mps, script uses CPU

Step 5: Verify & Report

Script automatically reports:

  • Output SRT file path
  • Number of subtitle entries
  • Processing time

Examples

Basic extraction:

User: 帮我提取这个视频的字幕
→ bun scripts/extract_subtitles.ts "video.mp4" en

With language specified:

User: Extract English subtitles from this video
→ bun scripts/extract_subtitles.ts "video.mp4" en

Chinese video:

User: 提取中文字幕
→ bun scripts/extract_subtitles.ts "video.mp4" zh

Long video with specific model:

User: 提取这个一小时视频的字幕
→ bun scripts/extract_subtitles.ts "video.mp4" en base

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.