S.E.P.I.A. Control HUB
Set up your Voice-Assistant Environment
Connections
Authentication Server
Assist Server
Teach Server
Chat Server
Mesh-Node Server
Load and Change Settings
Persistent settings are the settings stored in the properties file of the Assist-Server.
Write key-value pair to config. Will take effect after server reload.
Ad-hoc Toggles
Ad-hoc settings are settings that will not persist on server restart (e.g. for testing).
Changes some settings ad-hoc without writing to config file. Will reset on server-restart.
Database Functions
Modify database configuration. Use with CAUTION!
Create new user
By default you register a new user by setting his email (real or "fake", e.g. test@sepia.local) on the whitelist first.
Edit user
As admin you can change user roles and delete ANY user (but the core admin). Other users can delete ONLY themselves.
Smart Home HUB
Devices
Quick start
- Load HUB info from server or write new settings (more at: Core Settings)
- Register SEPIA Framework once if required (currently only: FHEM)
- Get the list of devices from smart home HUB
- Configure devices for SEPIA as needed
- See the INFO PAGE for detailed instructions
Test sentence
Remote action
Define custom answers for your services. NOTE: Only works if answer module is set to 'elasticsearch'.
SEPIA Speech-To-Text Server
The SEPIA client app can handle speech recognition in two different ways, either using platform specific engines (available in Android, Chrome, Edge, Safari, etc.)
or using your own, independent, offline speech recognition server that can be configured to use customized acoustic and language models.
To learn how to set up and use the server please visit:
SEPIA STT-Server @ GitHub
Server Connection
The STT server connection and model selection is handled directly inside your client app. Check out the menu at:
App Settings - General - ASR engine
ASR Model Adaptation
This feature is currently being redesigned to make it more accessible to users. In the meantime please follow these instructions:
Customize your ASR Models
To build your own language model you can start with a list of all the sentences you (or other users) have added to SEPIA. The Assist-Server has an endpoint to export this data:
Text-To-Speech (TTS)
Here you can test the integrated TTS module of SEPIA.
The priority of settings is 'mood - gender - language - voice' meaning for example
voice will always overwrite other selections.
Create channel
NOTE: When you create a channel as admin they will be added as a member of that channel. This may not be desired.
Channel history statistics
Get info about all (cached) channels that store a chat history. Check e.g. channel ID, size and last message poll.
Client Connections
Use this page to communicate with clients that are connected to a CLEXI server.