# Give your agents a voice with Smallest AI

> New SmallestTools lets an agent generate speech with Smallest AI's Lightning models. Call text_to_speech to turn text into natural audio, or get_voices to see what's on offer.

- Published: 2026-07-30
- Author: Agno Team
- Category: Changelog
- Canonical: https://agno-com-nine.vercel.app/articles/give-your-agents-a-voice-with-smallest-ai
- Markdown: https://agno-com-nine.vercel.app/articles/give-your-agents-a-voice-with-smallest-ai.md

New `SmallestTools` lets an agent generate speech with Smallest AI's Lightning models. Call `text_to_speech` to turn text into natural audio, or `get_voices` to see what's on offer. Nothing extra to install beyond `agno`, since the toolkit runs on Agno's own HTTP client, so it comes down to setting your `SMALLEST_API_KEY` and going.``

There are two tiers to work with. `lightning_v3.1` is the default and supports cloned voices, while `lightning_v3.1_pro` opens a premium voice pool with American, British, and Indian accents for broadcast-quality output. Set `target_directory` and the audio saves itself, or leave it off and pull the bytes off `response.audio` to pass along however you like. Output can come back as WAV, MP3, or a few telephony formats if you're feeding this into calls.

Check out [Agno’s Smallest AI toolkit docs](https://docs.agno.com/tools/toolkits/others/smallest#smallest-ai) for more.