# lalia

> The visual layer for voice AI: fluid, audio-reactive orbs for voice agents. Zero dependencies, raw WebGL2, React component or framework-free engine. Six states (idle, connecting, listening, thinking, speaking, muted) that glide into each other, driven by microphone and agent audio.

Install with `npm install lalia`. Entry points: `lalia` (engine, audio, states), `lalia/react` (`<Orb>`), `lalia/variants` (designs), `lalia/schemes` (colors), `lalia/adapters/{elevenlabs,openai,gemini}`.

## Docs

- [Full reference](https://lalia.offbrand.design/llms-full.txt): the complete docs in one Markdown file
- [Install prompt](https://lalia.offbrand.design/prompt.md): step-by-step instructions for adding the orb to an app
- [Agent skill](https://lalia.offbrand.design/SKILL.md): a skill file for Claude Code, Droid and other agents that load SKILL.md

## Designs

- Mist: wisp, haze, cloud
- Silk: silk, satin
- Glass: glass, prism
- Lumen: lumen
- Chrome: chrome

## Schemes

- Neutral: graphite, pearl, linen, glacier
- Hue: cobalt, iris, orchid, rose, coral, ember, honey, citrus, sage, jade, lagoon
- Atmosphere: dusk, dawn, borealis, nocturne, abyss, opal, cosmic

## Optional

- [Playground](https://lalia.offbrand.design/play): try every design, state, scheme and your microphone, then copy the code
- [Docs](https://lalia.offbrand.design/docs): the same reference, for humans