Your own digital voice-assistant


What is SEPIA?

The "A" in SEPIA stands for assistant. Think of it as a digital companion that helps you whenever you need it, not unlike Apple's Siri, Amazon's Alexa, Microsoft's Cortana or Google Assistant, but a very special, private, customizable and open version that really belongs to you!
In fact SEPIA is not only a ready-to-use, cross-platform app for iOS, Android and desktop browsers but also a complete open-source framework for DIY AI assistants. You want to build your own smart-speaker or home-assistant or you simply want to use a virtual assistant that respects your privacy? Then you found the right place to start :-)

You control your data

Backbone of the SEPIA-Framework is your own, secure web-server. It runs all the core-modules like the Elasticsearch database, the user-account management, the natural-language-understanding (NLU) and the chat-module. Additionally it connects the NLU to smart-services like reminders, to-do lists, news, IoT interfaces and many more while keeping your data in a safe, self-hosted environment. Thanks to Java the SEPIA back-end works on virtually any OS (Linux, Windows or Mac). It is optimized for single-board computers like the Raspberry Pi to run inside your private network but can also be scaled up and hosted as cloud-service. Learn more

SEPIA connected

You shape your assistant

SEPIA customize

SEPIA is extendable and can be improved and individualized in many ways. You can create your own commands on-the-fly and teach SEPIA new stuff using the app (iOS, Android or browser) or you can build complex, voice-controlled smart-services with the Java SDK. Everything from a simple answer to a multi-step dialog with cloud-service integration is possible. If that's still not enough feel free to dive deeper into the code and replace certain modules or rebuild SEPIA completely. Every part of the framework is open-source, free and everybody can contribute.
Learn about the APP Learn about the SDK

Be connected while staying private

Since you control the data and shape your assistant SEPIA's services and answers can use your personal profile to give you individual results without this strange feeling in your stomach that you might get when you don't know who else is reading your messages or collecting information about you. Connect your own IoT devices to SEPIA and keep all the controls protected in your own environment or use a third-party service like Google Maps but share only an anonymous address without any context so Google will never know if you are driving to your grandma or your girlfriend/boyfriend.

SEPIA personal

Natural dialog and smart actions

SEPIA smart

SEPIA is designed to understand requests spoken (or typed) in natural language, extract as much information as possible and combine it in an intelligent way with an ever growing number of useful services. If a service is missing some parameters to fulfil the request SEPIA can start a multi-step conversation and ask you questions to fill the gaps. On top of that SEPIA will try to show you relevant information depending for example on the time of the day or your personal preferences.

DIY voice-controlled devices

With SEPIA Mesh-Nodes you can extend the reach of your assistant and build your own voice-controlled device. All your device needs is network access (e.g. via WiFi module) and the ability to run the lightweight Mesh-Node server (e.g. via Java). Mesh-Nodes connect securely to SEPIA and can execute programs directly on your device. Via the SEPIA client app, you can create voice-commands for your Mesh-Node or you can use the SEPIA Control-HUB to code and upload programs. Learn more

SEPIA mesh-node

There are no limits to your creativity!

Privacy policy GitHubSource code Documentation TwitterTwitter BlogBlog Credits