TECHNOLOGY
Utilizing state of the art technology allows us to offer a Speech Recognition platform that strikes a perfect balance between accuracy, cost, and features.
All VoiceGain platform users will appreciate the high accuracy of our speech recognizer utilizing Deep Neural Network technology.
​
Cloud users will like the ease of use of the modern Web API, and On-Prem users will be able to manage their own deployments using just a few mouse clicks, thanks to underlying modern software architectures and tools.
APPLICATIONS
Speech Recognition applications come in many shapes and sizes. That is why our platform supports a variety of application scenarios:
-
web api for transcription - real-time and off-line
-
web api for speech recognition with context free grammars and semantic tagging,
-
legacy IVR support via MRCP interface
-
modern webhook-based IVR with SIP/RTP interface
DIFFERENCE
Here are just a few aspects that differentiate VoiceGain from other ARS products:
-
our ASR is available both in cloud and on-prem
-
our platform enables customization of the acoustic model - most others provide only language model customization
-
our pricing is very competitive
-
we go beyond bare bones web-API and provide methods and tools to ease implementation
-
our MRCP interface provides a unique drop-in replacements for ASR in legacy IVRs
KEY FEATURES
ACCURATE
ASR
Ultimately what matters in an ASR is the accuracy. Out of the box, our ASR has over 90-95% accuracy on a wide range of voices and domains. However, with very little effort it can be customized to achieve 98%+ accuracy on real life transcription tasks.
IVR
COMPATIBLE
Cloud ASRs accessible only via Web-API are not easy to use for typical IVRs as, e.g., used for Phone Customer Support.
​
VoiceGain ASR was designed from ground up with traditional IVRs in mind. It supports standard MRCP protocol, which makes it a drop in replacement for most VXML-based IVRs.
​
EASY
API
Modern restful web API allows for easy integration with customer applications. We provide various options for feeding the audio data into the ASR, including several options for real-time streaming. There are also several ways to retrieve the output of the ASR including websockets and HTTP/2.
ON-PREM
OPTION
While cloud-based APIs cannot be beat for convenience, many of our customers prefer to deploy our ASR platform on-prem. This can be done with just a few mouse clicks via our web portal.
​
On-prem has many advantages: full control over data and security; custom acoustic models; support for MRCP and RTP; lower price.
​
PRICING
IN CLOUD
1.20 cents
per minute
Off-Line
or Preemptible
ASR
1.60 cents
per minute
Real-Time
Non-Preemptible
ASR
ON-PREM
1.20 cents
per minute
Standard
Model
ASR
1.60 cents
per minute
User-Trained
Model
ASR
Large volume discounts are available to customers using OnPrem deployments.