Core server of the SEPIA Framework responsible for NLU, conversation, smart-service integration, user-accounts and more.
Branch: master
Clone or download
fquirin Merge pull request #14 from SEPIA-Framework/dev
Multiple adjustments for SDK and SEPIA-Home v2.2.0
Latest commit 8e7c1d5 Jan 31, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
src fixed an issue with "smart" assumptions in Alarms Jan 28, 2019
.gitignore changed version number to 2.1.5, javadoc pom entry and micro-fixes Dec 26, 2018
LICENSE added license and readme Jun 15, 2018
pom.xml fixed surefire maven plugin (for testing) Jan 28, 2019

SEPIA Assist-Server

Part of the SEPIA Framework

S.E.P.I.A. Framework

This is the core server of the SEPIA Framework and basically the "brain" of the assistant. It includes multiple modules and microservices exposed via the Assist-API, e.g.:

  • User-account management
  • Database integration (e.g. Elasticsearch)
  • Natural-Language-Understanding (NLU) and Named-Entity-Recognition (NER) (currently in german and english)
  • Conversation flow (aka interview-module)
  • Answer-module
  • Smart-services (integration of local-services like a to-do lists or cloud-services like a weather API with the NLU, conversation and answer modules)
  • Remote-actions (e.g. receive data from IoT devices or wake-word tools)
  • Third-party embedded and cloud Text-to-Speech integration (optional, TTS is primarily handled in the SEPIA client)
  • ... and more

The SEPIA cross-platform-clients can access the RESTful Assist-API directly and exchange data in JSON format (e.g. for user authentication) or connect to the SEPIA chat-server to send and receive messages. SEPIAs running on this server can log-in to the WebSocket chat-server the same way a user does and communicated via channels with multiple users (or devices) at the same time.

The SEPIA Assist-Server operates as your own cloud-service and is designed to work the same way no matter if you run it on a Raspberry Pi for a small group of users in a private network or when you host it on multiple servers for a larger company network.