Audio Transcribe

Converts speech to text with timestamps and optional speaker identification, which is exactly what you need for meeting transcriptions, subtitle generation, or any voice recording processing. Supports multiple providers: run locally with Transformers.js if you want zero API costs (though no speaker ID), or use fal, Replicate, or RunPod for cloud processing with diarization. The JSON output includes segment-level timing and speaker labels when enabled. One honest take: the local provider uses Moonshine instead of Whisper and claims 5x speed gains, but you'll need to hit the cloud providers if speaker identification matters for your use case. Works with mp3, wav, m4a, and ogg files, plus auto-detects language if you don't specify it.

skills.sh installs: 194
GitHub stars: 6
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add agntswrm/agent-media --skill audio-transcribe --agent claude-code

Installs into .claude/skills of the current project.

Files

Audio Transcribe

Transcribes audio files to text with timestamps. Supports automatic language detection, speaker identification (diarization), and outputs structured JSON with segment-level timing.

Command

npx agent-media@latest audio transcribe --in <path> [options]

Inputs

Option	Required	Description
`--in`	Yes	Input audio file path or URL (supports mp3, wav, m4a, ogg)
`--diarize`	No	Enable speaker identification
`--language`	No	Language code (auto-detected if not provided)
`--speakers`	No	Number of speakers hint for diarization
`--out`	No	Output path, filename or directory (default: ./)
`--provider`	No	Provider to use (local, fal, replicate, runpod)