Main component of solution, Voise Server is the gateway among the client and the many cognitive providers, such as speech recognition (Speech-to-text), voice synthesizer (Text-to-speech) and intention recognition.
Currently follow providers are supported:
Provedor | Tipo | STT | TTS | IR |
---|---|---|---|---|
Google Cloud | Saas | X | X | |
Microsoft Azure | Saas | X | ||
Microsoft Speech | On-premise | X | ||
Weka | On-premise | X |
- .Net Framework 4.8 Runtime (https://dotnet.microsoft.com/download/dotnet-framework/net48)
- If using the Microsoft Speech provider
- Microsoft Speech Platform Runtime - x64
- Language Packs SR (ou STT) and TTS of chosen languages
- Compile the project
- Move the artefacts to %ProgramFiles%/voise-server directory.
- Create the Windows service, executing the follow command (as admin mode):
cd %ProgramFiles%\voise-server
Voise.Server.exe install
The service is created with automatic start.
The configuration is done in config.xml file in solution's main directory.
Example of configuration using the Google Cloud:
<?xml version="1.0" encoding="utf-8" ?>
<config>
<!-- Default: 8102 -->
<!--<port>8102</port>-->
<recognizers>
<!-- Default: me -->
<enabled>ge</enabled>
<google>
<credential_path>.\credentials\google-credential.json</credential_path>
</google>
</recognizers>
<synthesizers>
<!-- Default: me -->
<!--<enabled>me</enabled>-->
</synthesizers>
<!-- Default: ./classifiers/ -->
<!-- <classifiers_path>./classifiers/</classifiers_path> -->
<!-- WARNING: The log configuration must be directly done at Log.config file -->
<tuning>
<!-- Default: false -->
<enabled>true</enabled>
<!-- Default: ./tunning/ -->
<!-- <path>./tunning/</path> -->
<!-- Default: 7 -->
<!-- <retention_days>7</retention_days> -->
</tuning>
</config>