speech-video-transcriber

turn a local media file into a markdown transcript with minimal local compute.

the intended path is:

• extract compact mono audio with ffmpeg
• send that audio to OpenAI transcription
• write a markdown transcript into ../transcriptions/ relative to the skills/ directory

for this repo, that means outputs land in:

• /Users/rami/Documents/life-os/ai-agents-config/transcriptions/

when to use it

use this skill when the user wants any of the following from a local media file: