VoiceGain architecture consist of the following functional components distributed over VoiceGain Cloud, Customer OnPrem/Edge setup (zero, one, or multiple) and Auxiliary Services located in the Cloud.
VoiceGain Cloud provides the following functions:
- Web API - mainly used to access ASR functionality via RESTful API
- Cloud Web Console - Provides means for managing the VoiceGain resources by Customer accounts
- IVR - either
- MRCP based with a MRCP proxy installed on Customer side
- (coming soon) IVR with call-back style functionality. Comes with SIP interface
- Storage - used to temporarily store results and input data under customer control
Edge (OnPrem) Customer install is deployed on top of Customer provided Kubernetes Cluster. Usually Customer will also provision Document, Object, and Log Storage.
Customer can manage their Edge Deployment via the Cloud Web Console.
Customers have full control over their data. VoiceGain has no access to Customer provisioned storage, except of course that VoiceGain software processes the audio data for recognition. All of this done on the Customer hardware infrastructure (whether on-premise or private cloud).
Edge Customer deployment provides the following functions:
- Web API - mainly used to access ASR functionality via RESTful API. Reachable only over local IPs.
- Local Web Console - Provides means for managing the the local VoiceGain platform install.
- IVR - Comes with SIP and MRCP interface
- Tools - these are mainly QA tools for ASR recognizer tuning.
- Storage - this is Customer Provisioned storage used to store all of the local data.
Other Cloud Services are used to provide:
- Allocating phone numbers and SIP trunks to IVR.
- Billing functionality and also Customer Billing Portal.
- Knowledge base and support pages.
- Bulk Storage used for deploying large data sets to various Customers installs.