speak

Convert AI text responses into audible speech via multiple TTS backends.

8|4|Updated Jan 11, 2026

Search Tags:#accessibility #audio #tts #speech synthesis #speaking #voice output

Ready to run

npx skills add https://github.com/alamparelli/mcp-claude-say --skill speak-alamparelli

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: speak
Source: https://github.com/alamparelli/mcp-claude-say/tree/main/skill
Command: npx skills add https://github.com/alamparelli/mcp-claude-say --skill speak-alamparelli

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill allows Claude to vocalize its responses, enabling users to listen to Claude's output instead of reading it, enhancing accessibility and providing an auditory interaction experience.

Core Features & Use Cases

Text-to-Speech (TTS): Converts Claude's text responses into spoken audio.
Multiple Voices & Languages: Supports various TTS backends (macOS, Kokoro, Google) with different voice options.
Expressive Modes: Allows for different speaking styles (brief, brainstorming, complete) to match the conversational context.
Use Case: When you're multitasking or prefer an auditory experience, you can have Claude read out summaries, explanations, or even engage in a spoken conversation.

Quick Start

Have Claude say "Hello, how can I help you today?" in a brief mode.

Frequently Asked Questions about speak

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

FAQPage Schema

How do I convert text responses into spoken audio using a text-to-speech tool?▼

Text-to-speech functionality converts written text into audible speech, allowing you to listen to AI responses. This Skill supports backends like macOS 'say', local MLX 'kokoro', and cloud-based 'google' TTS for vocalizing outputs.

Does text-to-speech output work with macOS and cloud-based TTS backends?▼

Yes, the text-to-speech output supports multiple backends including native macOS 'say', local MLX-based 'kokoro', and cloud-based 'google' TTS. This provides flexibility in choosing your preferred voice synthesis environment.

What's the best way to control speaking speed and voice selection for speech synthesis?▼

The best way to control speech synthesis is by using the built-in settings for speaking speed and voice selection. You can choose different TTS backends and distinct modes like brief, brainstorming, or complete explanations to match your context.

Can I use different speaking styles for brainstorming and brief explanations?▼

Yes, you can use different speaking styles for brainstorming and brief explanations. The Skill offers distinct expressive modes—brief, brainstorming, and complete—allowing the spoken audio to match the conversational context appropriately.

Why use text-to-speech for AI model interactions instead of reading text?▼

Text-to-speech for AI model interactions enhances accessibility and provides an auditory experience. It is useful when multitasking or when you prefer listening to summaries and explanations rather than reading text on a screen.

What are the limitations of local MLX-based speech synthesis compared to cloud TTS?▼

Local MLX-based 'kokoro' speech synthesis runs entirely on your machine, offering privacy but requiring local compute resources. In contrast, cloud-based 'google' TTS handles processing remotely, shifting compute load but requiring an internet connection.