Speech Recognition | ReconVox

ReconVox is our continuous speech recognition engine. It can not only get the transcription out of a free speech phone conversation, an IVR system or radio and TV broadcasts, but it can also work in Word-Spotting mode.

In this mode it’s possible to search into voice recordings (or detect in real time from a streaming audio) for specific keywords or short sentences defined on the fly, discarding all the surrounding context that could even be in a different language.

ReconVox exports all its functionality via an easy to use yet powerful API designed to be integrated into any on-premise system. A SaaS cloud solution is coming soon.

ReconVox is 100% developed by DTec and allows for a full customization from scratch of the vocabulary and language model. Available in English and Spanish, new languages under preparation.


Download the PDF with ReconVox’s Product Information



AutoLearn is a technology developed by DTec that allows continuous accuracy improvement over time. It can work in two modes:

  • Automatic: AutoLearn manages by itself all the learning process, the user just switches it on and keeps using ReconVox as usual. When eventually enough information is available, AutoLearn trains a set of improved acoustic models that are immediately used and a new learning cycle is started.

  • Supervised: if the learning process is to be optimized and accelerated, it’s possible for the user to provide a set of selected voice recordings together with their transcriptions. The user has full control of the quantity and type of the audio to be used in the learning iteration.

AutoLearn can be used for improving the accuracy for a specific speaker, but also for adapting to problematic dialectic regions or even noisy acoustic environments.

Automatic transcription:

  • Free speech with agents in call-centres.

  • IVR systems with open questions.

  • Radio and TV broadcasts.

Speech Analytics based on Word-Spotting:

  • Searching for keywords in voice recordings.

  • Real time keywords detection in audio streaming.

  • Media clipping (commercial brands and personalities).

Electronic devices, domotics and video games:

  • Alarms, domotics and electronic systems in the vehicle controlled by voice.

  • Communication with NPCs in video games or with your pit box engineer in racing games.

  • Voice control for inventory, maps, avionics in flight simulators…

  • Free speech answers in quiz games.

Language education:

  • Analysis and scoring for real of pronunciation in e-learning platforms.

  • Free speech answers in conversation exercises.