> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ruby.ad/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice and sound generation

To speak with an agent in real time and hear its replies, see [Ruby Live](https://docs.ruby.ad/docs/user-documentation/agents/ruby-live.md).

## Speech Generation

| Name | What the agent will do | How to configure it when building an agent |
| - | - | - |
| Speech Generator | Convert written text into natural-sounding audio with configurable voices, languages, and speaking styles. Generate single-voice narrations or multi-speaker dialogues for podcasts, training content, and audio communications. | No configuration required; simply enable this capability |

**Key capabilities:**

* **Single voice audio:** Generate up to 10,000 characters of speech with customizable gender, language (21 languages supported), and use case (conversational, narrative, educational, etc.)
* **Multi-speaker dialogue:** Create natural conversations between multiple speakers, supporting up to 5,000 dialogue lines for podcast-style content
* **High-quality output:** All audio generated in MP3 format at 44.1kHz, 128kbps for professional-grade results

**Common use cases:**

* Transform research from `@deep-dive` into podcast episodes
* Create audio versions of written content for accessibility
* Generate training materials and onboarding content in audio format
* Produce multilingual voice content for global teams


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.