An agent framework that connects digital humans (2.5D, 3D, mobile, PC, web) and LLMs (OpenAI-compatible, DeepSeek) to business systems.
SEPIA-Framework/sepia-html-client-app
Cross-platform browser, Android and iOS client for the SEPIA personal voice assistant, with chat, speech input and TTS output.
- Stars
- 65
- Forks
- 14
- Open issues
- 2
- Last commit
- 23 days ago
What is it?
The SEPIA app is the main client for talking to a SEPIA assistant. It is written in JavaScript and HTML, runs in any modern browser on desktop, Android and iOS as a web or PWA app, and can also run headless as a smart speaker or smart display on single-board computers such as the Raspberry Pi.
Users can control the assistant by voice, touch or text. Speech recognition uses native platform APIs or the open-source SEPIA STT server, and speech output uses native APIs, the SEPIA server or tools such as Mary-TTS and Larynx. Built-in wake-word support is provided through the SEPIA Web-Audio library, and an always-on mode shows an animated avatar.
The client also supports chat with other users on the same server over WebSocket, in-chat cards and audio playback, custom HTML voice and media-player widgets, a teach interface for adding commands, and remote actions between devices.
Key features
- 01Voice, touch and text control on Android, iOS and modern browsers
- 02Speech recognition via native APIs or the open-source SEPIA STT server
- 03Wake-word support and an always-on mode with animated avatar
- 04Chat with other users over a private WebSocket connection
- 05Custom HTML voice widgets and media-player widgets
- 06Teach interface for adding new commands and actions
Use cases
- Running a voice assistant on a Raspberry Pi as a headless smart speaker or smart display
- Using a self-hosted SEPIA assistant from a phone home screen as a PWA
- Building custom voice widgets or media players for the SEPIA client
- Teaching a personal assistant new commands through the teach interface
Quick start
- 1
Open the latest web app in your browser
- 2
Install the Android app from Google Play
- 3
On iOS, open the default link in Safari and add it to your home screen as a PWA
Things to know
- Many features require your own SEPIA server; the public demo mode offers only limited functionality.
- The native iOS app development is on hold, so iOS use relies on the PWA in Safari.
In collections
Topics
FAQ
- What is sepia-html-client-app?
- Cross-platform browser, Android and iOS client for the SEPIA personal voice assistant, with chat, speech input and TTS output. The SEPIA app is the main client for talking to a SEPIA assistant. It is written in JavaScript and HTML, runs in any modern browser on desktop, Android and iOS as a web or PWA app, and can also run headless as a smart speaker or smart display on single-board computers such as the Raspberry Pi.
- Is sepia-html-client-app free to use? What license does it have?
- sepia-html-client-app is open source under the MIT license. Check the license terms before using it in commercial products.
- Is sepia-html-client-app still maintained?
- Actively maintained. The last commit to sepia-html-client-app was on Sep 29, 2026.
- What can I use sepia-html-client-app for?
- Running a voice assistant on a Raspberry Pi as a headless smart speaker or smart display Using a self-hosted SEPIA assistant from a phone home screen as a PWA Building custom voice widgets or media players for the SEPIA client Teaching a personal assistant new commands through the teach interface
- How do I install sepia-html-client-app?
- Open the latest web app in your browser Install the Android app from Google Play On iOS, open the default link in Safari and add it to your home screen as a PWA
- Where can I get sepia-html-client-app?
- The source code and installation instructions are on GitHub: https://github.com/SEPIA-Framework/sepia-html-client-app
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Capacitor is a cross-platform runtime that lets web apps run natively on iOS, Android, and the Web from one codebase.
Self-hostable multi-agent personal AI assistant built on LangGraph, connecting mail, calendar, documents and tasks with human approval before actions run.
Java server and admin platform for Xiaozhi ESP32 voice devices, covering device monitoring, voice and persona management, and conversation logs.
An open-source, on-device voice assistant for Android with speech and graphical replies and a set of built-in skills.
Offline speech toolkit using onnxruntime for speech recognition, text-to-speech, VAD, diarization and more, across many platforms and languages.