skills.rest
fairchild
fairchildCOMMUNITY
Explore all skills by this author →
GitHub

vocal

Speak text aloud and transcribe speech using local macOS tools or ElevenLabs APIs.

1|Updated Nov 11, 2012
Search Tags:#text-to-speech#transcription#whisper#macos#elevenlabs#speech-to-text#voice assistant
One-click install
npx skills add https://github.com/fairchild/dotfiles --skill vocal-fairchild
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vocal
Source: https://github.com/fairchild/dotfiles/tree/main/agents/shared/first-party-skills/vocal
Command: npx skills add https://github.com/fairchild/dotfiles --skill vocal-fairchild

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires elevenlabs, httpx, mlx-whisper, sounddevice, numpy, and includes scripts (resource) and references (resource) components.

What problem does it solve? Voice interaction with an AI assistant normally requires manual typing and reading; this Skill adds text-to-speech and speech-to-text capabilities so you can run a hands-free, turn-based voice conversation loop or transcribe audio files. ## Core Features & Use Cases - Turn-based vocal loop: Run /vocal to start an ask-aloud / listen / respond cycle driven by a background listener agent, with configurable STT/TTS providers and recording duration. - Multiple providers: Use local macOS say and mlx-whisper (free, offline) or ElevenLabs Flash/Scribe cloud APIs (low latency, high quality) interchangeably. - Web tuning console: Launch a local browser console at http://127.0.0.1:8765 to compare voices, test providers, record microphone audio, and save preferences. - Use Case: While away from the keyboard, start /vocal stt=local tts=local duration=8 What's next? and hold a spoken conversation where each reply is read aloud. ## Quick Start Ask the assistant to start vocal mode with a spoken prompt, for example by invoking /vocal with the text you want spoken aloud first.

Frequently Asked Questions about vocal

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

FAQPage Schema
How do I transcribe speech to text on macOS with Python?▼

Run the stt_local.py script with --duration to record from the microphone or --file to transcribe an existing audio file. It uses mlx-whisper on Apple Silicon and prints the transcript to stdout.

How do I use ElevenLabs for text-to-speech from the command line?▼

Set ELEVENLABS_API_KEY in ~/.env or your shell, then run tts_elevenlabs.py with --text and optionally --voice and --model. It generates an mp3 and can play it immediately with --play.

Does local speech transcription work without an API key?▼

Yes, local STT uses mlx-whisper and local TTS uses the built-in macOS say command, both free with no API key. Local STT requires Apple Silicon; the first run downloads the Whisper model weights.

Why does microphone transcription fail with permission errors?▼

macOS blocks microphone access until granted. Open System Settings -> Privacy & Security -> Microphone, allow Terminal or your host app, then re-run the command.

Is the vocal loop real-time full-duplex conversation?▼

No, the /vocal loop is turn-based, not full-duplex realtime. Each listen cycle is a separate background agent turn, so expect pauses between speaking and hearing the response.

Related Skills

Explore More Software Engineering Skills→
vellum-aivellum-ai
1.0k

voice-setup

Configure push-to-talk, microphone permissions, and ElevenLabs TTS in Vellum.

Official
Intermediate
DavidROliverBADavidROliverBA
49

voice-chat

Transcribe spoken commands into Claude Code and synthesize responses with Inworld TTS.

Community
Intermediate
yoonhwanyoonhwan
3

yt-transcribe

Transcribes and summarizes YouTube videos using local speech-to-text tools.

Community
Intermediate
ciniencecinience
396

alicloud-ai-audio-asr-realtime

Transcribe streaming microphone audio with Alibaba Cloud Qwen ASR Realtime models.

Community
Intermediate