A streamlined CLI tool that brings ElevenLabs generative text-to-speech to AI workflows with a macOS-style say interface.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install miranda-sag
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install miranda-sag using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is sag?
sag is a high-performance command-line interface designed to bridge the gap between ElevenLabs' state-of-the-art generative voice technology and the simplicity of the classic macOS say command. It allows developers and AI agents to generate expressive, human-like speech with minimal friction, supporting advanced features like emotion tagging, model selection, and local playback.
By integrating this tool into your Openclaw Skills library, you can empower AI agents to communicate through audio dynamically. Whether you need the nuanced expression of the v3 model for character-driven narration or the low-latency flash_v2_5 model for real-time interactions, sag provides the flexibility and control required for professional-grade voice synthesis.
sag Use Cases
- Automated voice replies for AI chat agents to provide immersive user experiences.
- Local playback of text content for accessibility or convenience during development.
- Generating high-quality audio assets for tutorials, videos, or notification systems.
- Testing ElevenLabs voice settings and model responses via CLI before production deployment.
- Creating character-specific voices like a crazy scientist or a calm narrator using emotional tags.
How sag Works
- The user or agent provides text input to the sag command along with optional parameters like voice ID or model type.
- sag processes the input, applying normalization rules or emotional markers such as [whispers] or [excited].
- The tool authenticates with the ElevenLabs API using a configured API key to synthesize the requested audio.
- Depending on the flags used, the audio is either played back locally or exported as a high-quality audio file.
- AI agents can then use the generated file path to serve audio responses back to the end user.
sag Setup
To install sag using Homebrew, execute the following command:
brew install steipete/tap/sag
Next, configure your environment with your ElevenLabs API key by adding it to your shell profile:
export ELEVENLABS_API_KEY="your_api_key_here"
You can verify the installation by running a quick test:
sag "Hello world"
sag Data Schema & Taxonomy
sag utilizes environment variables for configuration and can export audio data to the local filesystem. Below are the primary data components:
| Attribute | Description |
|---|---|
| ELEVENLABS_API_KEY | The primary authentication key for the ElevenLabs API. |
| ELEVENLABS_VOICE_ID | The default voice identity used for synthesis (e.g., Clawd). |
| Models | Supports eleven_v3 (expressive), eleven_multilingual_v2 (stable), and eleven_flash_v2_5 (fast). |
| File Output | Generates audio files, typically in .mp3 format, when the -o flag is specified. |
sag Advanced Features
- Expressive Audio Tags: Support for inline markers like [whispers], [shouts], [sings], and [laughs] when using v3 models.
- Normalization Control: Use the --normalize flag to manage how numbers, units, and URLs are processed.
- Language Guiding: The --lang flag allows you to guide the normalization process for specific languages like German, French, or English.
- Customizable Pacing: Implement dramatic timing using [pause], [short pause], or [long pause] to mimic natural human speech patterns.
- SSML Support: Full compatibility with standard SSML break tags for v2 and v2.5 models.
SKILL.md
Loading
Related Openclaw Skills
ElevenLabs Speech & Scribe
A dual-purpose voice solution providing high-fidelity text-to-speech synthesis and accurate speech-to-text transcription powered by ElevenLabs.
soul-md-maker
A comprehensive personality builder for OpenClaw agents that generates production-ready SOUL.md files through guided interviews, blending, or a pre-built gallery.
OpenClaw Email Lead Generation
A multi-tiered email outreach and pipeline management system for AI agents that automates lead tracking, sequence generation, and sentiment-aware follow-ups.
OpenClaw Cost Optimizer
A sophisticated cost-management skill that slashes API expenses by dynamically advising on the most cost-effective model for every task.
Devinism
A benign memetic experiment and AI safety demonstration that uses the concept of an AI religion to highlight information propagation in autonomous Openclaw Skills.
Clawvicular
Clawvicular automates the delivery of trending looksmaxxing slang and real-time influencer news using an authentic, extremely online community voice.
Featured*
