skills.rest
jterratsdev
jterratsdevOFFICIAL
Explore all skills by this author →
WebsiteGitHub

transcription-evidence

Transcribe audio and video files and record transcripts as task evidence.

Updated Dec 1, 2025
Search Tags:#video#transcription#whisper#evidence#audio#ffmpeg#speech-to-text
One-click install
npx skills add https://github.com/jterratsdev/smart-deployment --skill transcription-evidence-jterratsdev
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: transcription-evidence
Source: https://github.com/jterratsdev/smart-deployment/tree/main/.setup-agents/skills/transcription-evidence
Command: npx skills add https://github.com/jterratsdev/smart-deployment --skill transcription-evidence-jterratsdev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires ffmpeg, whisper-cpp, insanely-fast-whisper, faster-whisper, openai-whisper, and includes scripts (resource) components.

What problem does it solve? Converting audio and video recordings into text transcripts and attaching them to task workflows is manual and error-prone. This Skill automates transcription across multiple backends and records the result as structured evidence tied to a task ID. ## Core Features & Use Cases - Multi-backend transcription: Automatically detects and uses whisper.cpp, insanely-fast-whisper, faster-whisper, openai-whisper, or cloud APIs (OpenAI, Deepgram, AssemblyAI) based on availability. - Video audio extraction: Extracts audio tracks from mp4, webm, and mkv files using ffmpeg before transcription. - Evidence recording: Attaches transcript summaries to a task ID with a role label via the evidence CLI, with a transcription-pending fallback when no backend exists. - Use Case: A QA engineer records a bug walkthrough video, then runs the script to transcribe it and attach the transcript as evidence to the corresponding task for audit purposes. ## Quick Start Transcribe the file demo-recording.mp4 and attach the transcript as evidence to task TASK-123 with the qa role.

Frequently Asked Questions about transcription-evidence

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I transcribe an audio or video file from the command line?▼

Run the transcribe.sh script with the file path and a task ID, optionally adding a role. The script detects an available backend such as whisper.cpp, faster-whisper, or openai-whisper, transcribes the file, and records the transcript as task evidence.

What audio and video formats are supported for transcription?▼

Supported audio formats are mp3, wav, m4a, ogg, and flac. Video formats mp4, webm, and mkv are supported by extracting the audio track with ffmpeg before transcription. Files must be under 200MB.

Which Whisper backend should I use for transcription?▼

whisper.cpp is checked first as the fastest local option, followed by insanely-fast-whisper for GPU, faster-whisper for CPU, and openai-whisper as a fallback. Cloud backends like OpenAI, Deepgram, or AssemblyAI are opt-in via the TRANSCRIBE_CLOUD_BACKEND variable.

Can I transcribe video files without installing ffmpeg?▼

No, ffmpeg is required for video files because the script extracts the audio track to a 16kHz mono WAV before transcription. Audio-only files like mp3 or wav do not require ffmpeg.

What happens if no transcription backend is available?▼

The script exits with code 3 and records a transcription-pending evidence entry against the task ID if the evidence CLI is available. You can then install a local backend or set TRANSCRIBE_CLOUD_BACKEND with the matching API key.

Related Skills

Explore More Software Engineering Skills→
atlasgroupczatlasgroupcz
2

video-analyze

Transcribe and analyze video and audio files or YouTube URLs using Gemini multimodal models.

Official
Intermediate
Animism001Animism001

youtube-transcript

Fetch YouTube video transcripts and metadata as JSON with language selection.

Community
Intermediate
JigsawStackJigsawStack
4

speech-to-text

Convert spoken audio into searchable transcripts with timestamps and speaker labels.

Official
Intermediate
bomberoxenviosdosruedasbomberoxenviosdosruedas

agency-voice-ai-integration-engineer

Build speech transcription pipelines using Whisper models and cloud ASR services.

Community
Advanced