FueTrek Co., Ltd. provides a better and faster way to communicate through with speech recognition, speech dialogue, and speech translation.

Products and Services - A one-stop shop company of speech recognition, speech synthesis, dialogue, and machine translation solutions

Inspection/inventory management Consumer electronics/home appliances Finance/insurance Automobile Contact Center Manufacturing Entertainment E-commerce Education Mobile Device UI

FueTrek creates value in a variety of settings, from speech recognition to speech dialogue, and electronic help solutions, sound generator IP, and memory card writing.

vGate: Speech-related Solution

FueTrek provides vGate as a speech-related technology suite based on world-class speech language technologies. In addition to standalone speech recognition software, we provide vGate ASR (speech recognition) and vGate TTS (speech synthesis), as well as vGate Talk2Me (speech dialogue), which combines vGate ASR and TTS, to create a new, more convenient user interface. Further, our vGate suite can increase efficiency and productivity in business, as well as increased sales.

The following vGate systems and services are available to help you achieve the optimal use of speech-related technologies in various business types and environments.

vGate Products and Services


vGate ASR Speech Recognition System

  • Server-type
    - Stores large lexicons and models on a server, allowing free-word text input and high-performance recognition.
    - Easy to add lexicons to be recognized and perform other maintenance.
  • Local-type
    - Recognition and conversion to text are carried out without the need for a network connection.
    - Optimal for pre-programmed phrases, terms, and command input consisting of tens to tens of thousands of sets of vocabulary words and models, such as data stored on local drives of devices.

vGate Talk2Me Speech Dialogue System

  • Scenario-type
    Suitable for usage in routine work by pre-programming user utterances and system responses to them as speech dialogue scenarios.
  • Inference-type
    Capable of interpreting the intent of user utterances and returning responses that match that intent, thereby providing users of various tastes and needs with the most appropriate information and recommending them the best suited products and services.

Speech Translation System

FueTrek's "vGate Translate" consists of FueTrek's "vGate ASR" (Automatic Speech Recognition) system, "vGate TTS (Japanese)" (Text to Speech), and machine translation technology of its subsidiary, ATR-Trek. "vGate Translate" covers wide range of use cases, not limited to tourist information services, ordinary conversation tools, but also a business usage such as communication tool at a manufacturing site.

Mechanism of Speech Translation System


Echo Cancellation

A speech-recognition-compatible echo cancellation product based on Artifit Voice by ARI Co., Ltd. It provides small-size operation, high speed convergence, and robustness simultaneously, with the usage of Fast H∞ filter (J-FHF)*, which was very difficult to achieve with conventional technologies. We have used the filter’s characteristic strengths against doubletalk and environmental noise to adapt it to a speech recognition barge-in function.

* "Fast H∞ filter (J-FHF)" is a technology developed by Prof. Kiyoshi Nishiyama of Iwate University in Japan, and The Japan Science and Technology Agency has applied for a patent for this technology in accordance with the patent acquisition system in effect in the applicable country.


  • Speech Recognition ASP Service


  • Voice Memo (Speech input of notes and simple editing function)
  • Voice UI
    - The standard speech input UI can be triggered in various applications.
    - In addition to interfacing as an independent speech recognition app, libraries are also available for embedding in applications.


  • Dictation Editor (Transcription software)

The Advantages of vGate Suite

The world’s top-class speech recognition engine*1 developed through joint research with ATR and NICT

vGate utilizes the results of joint research and element technologies developed through a strong partnership with speech-technology research pioneer Advanced Telecommunications Research Institute International (ATR) and the National Institute of Information and Communications Technology (NICT).

Because we have all the speech-related technologies needed for speech recognition, synthesis, dialogue, and machine translation, the FueTrek Group alone can provide voice agent services through the use of speech dialogue. Additionally, collaboration within the Group makes it possible to provide everything from developing apps to providing servers and network services.

Our speech-related technologies are available in Chinese (Mandarin), English (North America, UK), Indonesian, and Thai. Support for other languages is planned for the future.

*1 Refers to our Japanese-language speech recognition technology, based on our research.

Characteristics of FueTrek Speech Recognition Engine

FueTrek has a speech recognition engine with a high recognition accuracy, composed of noise-reduction technology that is effective at reducing non-stationary noise, statistical acoustic models created from speech data from a large number of speakers, language models collected from large corpora, and speech dictionaries that list the spelling of vocabulary. This allows selection of deployment methods according to the characteristics of the device and the specific use cases.
In addition, we continuously strive to improve the performance of our advanced acoustic models by conducting daily mass-scale learning of more than 500,000 people total, by operating various commercial services using our technologies.

Questions? Get in touch with us.Contact Us