AGENT SKILL

video-to-transcript

Use when the user wants to convert a local video or meeting recording into a text transcript, especially MP4/MOV recordings. Covers extracting audio with ffmpeg and transcribing locally with faster-whisper.

License 许可
MIT
所属仓库
srafis/skills
最近更新
2026-10-16
Stars
★ 2
Forks
0

01 前言 Front Matter

name
video-to-transcript
description
Use when the user wants to convert a local video or meeting recording into a text transcript, especially MP4/MOV recordings. Covers extracting audio with ffmpeg and transcribing locally with faster-whisper.

03 SKILL.md 正文

Video To Transcript

Use this workflow to convert a local video file into a timestamped text transcript.

Prerequisites

Check that ffmpeg and Python are available:

which ffmpeg
which python3

If Python package installation is needed, use a project-local virtual environment rather than system Python.

Workflow

  1. Create a working folder.
mkdir -p transcript_output
  1. Extract the audio from the video as mono 16 kHz WAV.
ffmpeg -y -i "/absolute/path/to/video.mp4" \
  -vn -ac 1 -ar 16000 -c:a pcm_s16le \
  transcript_output/audio.wav

Use the exact input video path provided by the user. The output WAV is a temporary helper file.

  1. Create and prepare a local Python virtual environment.
python3 -m venv .venv_transcript
.venv_transcript/bin/pip install --quiet faster-whisper

If the system Python is externally managed, do not use --break-system-packages; the local venv avoids that problem.

  1. Transcribe with faster-whisper.
.venv_transcript/bin/python - <<'PY'
from faster_whisper import WhisperModel
from pathlib import Path

audio_path = "transcript_output/audio.wav"
out_path = Path("transcript_output/transcript.txt")

model = WhisperModel("small", device="cpu", compute_type="int8")
segments, info = model.transcribe(
    audio_path,
    beam_size=5,
    vad_filter=True,
)

with out_path.open("w", encoding="utf-8") as f:
    f.write(f"Language: {info.language} probability={info.language_probability}\n\n")
    for segment in segments:
        line = f"[{segment.start:07.2f} - {segment.end:07.2f}] {segment.text.strip()}\n"
        print(line, end="")
        f.write(line)

print(f"\nWROTE {out_path}")
PY

Model choice:

  • Use small for a good balance of speed and accuracy on normal meeting recordings.
  • Use tiny or base if CPU transcription is too slow.
  • Use medium only when accuracy is more important than runtime.
  1. Verify the transcript.
sed -n '1,80p' transcript_output/transcript.txt
wc -l transcript_output/transcript.txt

Check that timestamps progress normally and that the detected language makes sense.

  1. Clean up helper files when the final transcript is no longer needed locally.
rm -rf transcript_output/audio.wav .venv_transcript

Only delete transcript_output/transcript.txt if the user has confirmed it is no longer needed.

Notes

  • For long recordings, the first run can take several minutes because the speech model may need to download and load.
  • If the transcript is used to create another artifact, keep the transcript until the artifact is accepted or the user asks for cleanup.
  • If the video has no audio stream, inspect it with:
ffprobe -v error -show_entries stream=index,codec_type,codec_name \
  -of default=nw=1 "/absolute/path/to/video.mp4"

04 同仓库更多 Skills

srafis
srafis/skills

work

Runs an Azure ticket delivery loop: fetch tickets, plan against design docs, implement via sub-agents, record before/after videos, run quality gates and CodeRabbit, then post ticket comments and commit only when asked. Use when the user says /work, names an Azure ticket or user story, or asks for a before/after video or ticket comment.

★ 2
srafis
srafis/skills

ui-ux-pro-max

UI/UX design intelligence for web and mobile. Searchable local database with 84 styles, 192 color palettes, 74 font pairings, 192 product types, 98 UX guidelines, 104 icon entries, 16 GSAP motion presets, and 25 chart types across 22 stacks (React, Next.js, Vue, Nuxt, Svelte, Astro, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, Jetpack Compose, Angular, Laravel, JavaFX, WPF, WinUI, Avalonia, Uno Platform, UWP, Three.js, and HTML/CSS). Use when designing, building, or reviewing UI: pages, components, color schemes, typography, layout, accessibility, animation, or data visualization.

★ 2
srafis
srafis/skills

ship-work

Use when shipping a software task end to end, with scope control and verified completion.

★ 2
srafis
srafis/skills

quick

Delegate task to fast model

★ 2
srafis
srafis/skills

meeting-mom

Use when the user wants to create Minutes of Meeting

★ 2
srafis
srafis/skills

humanizer

Rewrite AI-sounding text so it reads like the writer without changing what it says. Use when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line closers, staged openers, forced triads, dashes everywhere, inflated claims, sales language, stock AI words, bold labels, or filler. Based on Wikipedia's "Signs of AI writing."

★ 2