Can I use different ElevenLabs models?

Yes, sag supports eleven_v3 for expressiveness, eleven_multilingual_v2 for stability, and eleven_flash_v2_5 for speed.

How does Claude send audio files in a chat?

Claude uses the sag command to generate an .mp3 file to a local path and then includes a MEDIA: tag in its response for the system to render.

What API key is required for sag?

You need an ElevenLabs API key, which can be stored as ELEVENLABS_API_KEY or SAG_API_KEY in your environment variables.

How do I install the sag CLI?

You can install sag on macOS via Homebrew using the formula 'steipete/tap/sag' to enable the binary for Claude.

Does sag support emotional cues?

Yes, when using the ElevenLabs v3 model, you can use tags like [whispers], [shouts], [laughs], and [sarcastic] to guide the delivery.

ElevenLabs TTS (sag)

Name: ElevenLabs TTS (sag)
Author: openxjarvis

byopenxjarvis

•

Productivity & Workflow

Integrates high-quality ElevenLabs text-to-speech into Claude with a macOS-style command interface for local playback and expressive audio generation.

The sag skill enables Claude to generate and play back natural-sounding audio using ElevenLabs' advanced text-to-speech models. It mimics the familiar macOS 'say' utility, allowing users to specify voices, emotional tags like [whispers] or [excited], and custom pronunciation rules. This skill is particularly useful for developers wanting to add vocal responses to their AI interactions, preview audio content, or create character-specific voiceovers directly from the CLI environment.

Key Features

0140 GitHub stars

02Emotional expression tags including [whispers], [shouts], [laughs], and [sings]

03MacOS-style 'say' command parity for intuitive local audio playback

04Access to ElevenLabs v3, v2, and v2.5 models for expressive or low-latency TTS

05Seamless media integration for generating and linking audio files in chat responses

06Custom pronunciation controls with phonetic respelling and automated normalization

Use Cases

01Generating character-specific voice responses for interactive AI assistants

02Reviewing and debugging ElevenLabs TTS output with different models and parameters

03Adding accessible audio notifications to CLI-based developer workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add openxjarvis/openclaw-python sag

For use in Claude.ai and ChatGPT

Download Skill