Olivia

Settings

Backup & Restore

Downloads a complete backup of all your Olivia data as a JSON file.

Last Backup

Never

Restore from a previously exported Olivia backup file (.json).

Assistant Settings

Assistant

Assistant Name Shown in the AI Assistants sidebar list and chat header

Profile Picture

PNG, JPG, JPEG, WEBP — auto-resized to 256×256

Connection Status

Disconnected

Cannot delete the last remaining assistant

Connection

WebSocket URL Official server: wss://api.xiaozhi.me/xiaozhi/v1/

OTA / Provisioning URL Device registration endpoint (same as ESP32 firmware)

Pairing Status

Not paired

Devices pair automatically via 6-digit code at xiaozhi.me

Virtual Device Identity

Device-Id (MAC Address) Leave blank to auto-generate persistent MAC-style ID

Client-Id (UUID) Leave blank to auto-generate persistent UUID

Protocol Settings

Protocol Version

Frame Duration (ms)

Listening Mode

Audio

Enable Microphone (voice mode)

Play TTS audio from server

IDLE

Olivia initialized. Click Connect to register and pair via xiaozhi.me.

Connect to server to start chatting

Olivia (Open Language Intelligence & Voice Interactive Assistant) is a browser-based multi-assistant AI platform built on the Xiaozhi protocol.

Each assistant maintains its own independent identity, pairing, conversation history, connection, avatar, volume, and settings while sharing one unified interface.

Internally, Olivia emulates multiple virtual ESP32 devices, but this implementation detail is abstracted away to provide a seamless multi-assistant experience.

Features

Multi-assistant architecture

Independent assistant sessions

Independent conversations

Independent Xiaozhi pairing

Independent connections

Per-assistant avatars

Per-assistant volume controls

Light & Dark theme support

Persistent local storage

Browser-based platform

Cloudflare Pages deployment

Technology Stack

Frontend

TypeScript
Vanilla JavaScript
HTML5
CSS3

Backend

Hono
Cloudflare Pages
Cloudflare Workers

Communication

WebSocket
Opus Audio
Xiaozhi Protocol v1

Protocol Summary

TransportWebSocket (ws / wss)

Audio CodecOpus @ 16 kHz Mono

Provisioning6-digit Pairing Code

AuthenticationBearer Token

Protocol Version1

Assistant Architecture

Each assistant maintains its own independent:

Device ID
Client ID
Pairing Token
WebSocket Session
Conversation History
Avatar
Audio Volume
Settings

Internally, every assistant behaves as its own virtual ESP32 device, while presenting one unified application to the user.

System Status

Application VersionOlivia 2.2

Registered Assistants—

Connected Assistants—

Stored Conversations—

Theme—

StorageBrowser Local Storage

Open Source

Project Repository Documentation Issues

Version

Version2.2

ProjectOpen Language Intelligence & Voice Interactive Assistant

LicenseMIT

AuthorRoalf Burgonio

Initializing Olivia...

Settings

Backup & Restore

Import Backup

Assistant Settings

Assistant

Connection

Virtual Device Identity

Protocol Settings

Audio

OLIVIA

Protocol Debug Console

Features

Technology Stack

Protocol Summary

Assistant Architecture

System Status

Open Source

Version

Activation Required

Settings

Backup & Restore

Import Backup

Assistant Settings

Assistant

Connection

Virtual Device Identity

Protocol Settings

Audio

Take Photo

Protocol Debug Console

Features

Technology Stack

Protocol Summary

Assistant Architecture

System Status

Open Source

Version

Activation Required