gemini-media

name

gemini-media

description

Claude has no native audio or video understanding — this skill delegates that to Google Gemini 3.1 Flash Lite (cheap multimodal) on the ofox gateway. Use when the user asks to analyze / transcribe / describe / understand a sound, song, voice memo, or video, or asks what is said, sung, shown, or played in a media file. Triggers: "analyze this audio", "transcribe the song/speech", "what does this voice say", "describe/understand this video/mp4/clip", "what's the music/story in this MV", "listen to", "watch this".

gemini-media

Claude cannot hear audio or watch video. These scripts send the media to Gemini on ofox and return Gemini's answer as plain text, so an agent can reason over the result. Cost is a few cents; see the table below.

Scripts

Script	Purpose	Input
analyze_audio.py	Transcribe / describe sound, songs, speech	any audio file
analyze_video.py	Describe frames and audio of a video	any video file
fetch_bilibili.py	Companion: download a bilibili video + metadata to feed the analyzers	a BV id

Each is a standalone uv run script (PEP 723 inline deps). Requires ffmpeg and ffprobe on PATH, and the ofox API key in $OFOX_API_KEY.

Usage

# Audio — default prompt gives transcription + genre/mood, or ask your own
uv run analyze_audio.py song.mp3
uv run analyze_audio.py memo.m4a -p "Transcribe the speech, then summarize."

# Video — Gemini sees frames + hears the track
uv run analyze_video.py clip.mp4 -p 


uv run fetch_bilibili.py BV1ZA64BiEV2 -o mv.mp4
uv run analyze_video.py mv.mp4