skills.rest
Shakudo-io
Shakudo-ioOFFICIAL
Explore all skills by this author →
WebsiteGitHub

cartesia-tts

Generate real-time speech with the Sonic-3 model via WebSocket, SSE, and HTTP endpoints.

Updated Feb 14, 2026
Search Tags:#real-time#streaming#emotion#tts#telephony#voices#cartesia
One-click install
npx skills add https://github.com/Shakudo-io/opencode-skills --skill cartesia-tts
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cartesia-tts
Source: https://github.com/Shakudo-io/opencode-skills/tree/main/cartesia-tts
Command: npx skills add https://github.com/Shakudo-io/opencode-skills --skill cartesia-tts

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides real-time, natural-sounding text-to-speech with ultra-low latency for live voice applications, enabling responsive conversational AI, telephony, and embedded assistants across languages.

Core Features & Use Cases

  • Ultra-low latency TTS with Sonic-3 (~90ms first-byte) for real-time conversations, telephony, and Pipecat pipelines.
  • Streaming and batch endpoints: WebSocket, SSE, and HTTP; supports a multilingual voice library with 40+ languages and 60+ emotion controls.
  • Multilingual support and emotion controls suitable for customer service, IVR, voice assistants, and live narration.
  • Seamless integration with Pipecat workflows and telephony platforms, enabling real-time voice in AI pipelines.
  • Secure API usage with API keys and versioning; flexible output formats (raw PCM, WAV, MP3) and sample rates.

Quick Start

Send a text to the Cartesia TTS API using the sonic-3 model, select a voice, and request streaming audio in your desired format.

Frequently Asked Questions about cartesia-tts

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

FAQPage Schema
How do I add real-time text-to-speech to a conversational AI pipeline?▼

Real-time text-to-speech for conversational AI pipelines is provided via the Sonic-3 model with ultra-low latency of approximately 90ms to the first byte. It supports WebSocket, SSE, and HTTP streaming endpoints for immediate audio delivery.

Does real-time TTS support telephony platforms and voice assistants?▼

Yes, real-time TTS supports telephony platforms and voice assistants with flexible output formats including raw PCM, WAV, and MP3. It offers configurable sample rates to ensure compatibility with telephony and IVR systems.

Can I control emotion and language in streaming voice synthesis?▼

Streaming voice synthesis supports 40+ languages and 60+ emotion controls. This enables dynamic voice modulation for customer service, live narration, and multilingual conversational AI applications.

What is the best way to integrate text-to-speech with Pipecat workflows?▼

The best way to integrate TTS with Pipecat workflows is using the provided streaming endpoints alongside secure API key authentication. It seamlessly connects real-time voice synthesis into Pipecat AI pipelines for live audio processing.

What output formats are available for low latency voice synthesis?▼

Low latency voice synthesis outputs raw PCM, WAV, and MP3 formats. You can specify multiple sample rates to match your specific streaming pipeline or telephony platform requirements.

Related Skills

Explore More Software Engineering Skills→
Shakudo-ioShakudo-io

deepgram-stt

Convert spoken language into searchable text via the Nova-3 model.

Official
Advanced
WarrenZhu050413WarrenZhu050413
7

generating-tts

Generate and play multilingual text-to-speech audio with adjustable voices and speeds.

Community
Intermediate
ricardocidalericardocidale

marcela-ai

Configure a multi-channel AI assistant with ElevenLabs, Twilio, and RAG.

Community
Advanced
Shakudo-ioShakudo-io

elevenlabs-voice

Generate speech and sound effects via the ElevenLabs API.

Official
Advanced