Skip to main content

Service Specifications

This section summarizes the service specifications of AmiVoice API.

Security and Compliance​

Data Center Location and Data Processing​

In this service, speech recognition processing and data storage are performed in AWS data centers within Japan. Data will not be transferred, processed, or stored on servers outside of Japan.

Standard Specifications and Third-Party Certifications​

This service is operated under a management system based on the following standard specifications and third-party certifications:

About Credit Card Payment Information​

AmiVoice API uses Stripe's service for credit card payments.

Card information such as the card number and security code entered on the payment screen is obtained and processed directly by Stripe. It does not pass through our servers, and we do not store it.

Handling of Voice and Speech Recognition Result Data by Interface​

Voice data received for processing speech recognition requests and speech recognition result data are handled differently for each interface as follows. Please also see "Logging Options and Data Provision".

InterfaceData Location and Processing in Speech RecognitionRetention Period (*1)
Synchronous HTTP / WebSocketProcessed only in server memoryImmediately deleted upon completion of speech recognition processing
Asynchronous HTTPTemporarily stored on encrypted disk and processed in a distributed mannerDeleted after speech recognition processing is completed (*2)
  • *1 In case of "No logging". For "Logging", please see "In case of logging" in the next section.
  • *2 In the case of Asynchronous HTTP, the recognition results (text) are retained in an encrypted DB for 7 days and then automatically deleted. The recognition results stored in the DB can only be retrieved by the authenticated user.

Handling of User Dictionary Data​

There are two methods for registering user dictionaries, and the data is handled as follows for each method. Please also see "Logging Options and Data Provision".

MethodRequest ParameterOverviewStorage Location
Specified for each requestprofileWordsDirectly specify word data with the profileWords parameter at the time of request and perform speech recognition processing.Not saved
Specifying a pre-registered profileprofileIdSave to a profile in advance using the user dictionary registration API or MyPage. Specify the profile with the profileId parameter at the time of request and perform speech recognition processing.Encrypted storage

Note that user dictionary data is not used for learning at all.

Logging Options and Data Provision​

You can choose whether to provide data by setting the Logging option when making a speech recognition request.

In case of no logging​

It is not used for learning at all, and voice data and speech recognition result text are not recorded on the speech recognition server. Also, speech recognition results and user dictionary information set in the request parameter profileWords are not recorded in the operation log of the speech recognition server.

In case of logging​

A discounted price will be applied if you agree to provide data.

  • If logging is specified, after the speech recognition processing is completed, the voice data and speech recognition result data are copied to a dedicated encrypted storage.
  • We may use the voice data and speech recognition result data for research and development of our products and services, and for quality improvement. In principle, data used for learning for quality improvement will not be deleted, while data not used for learning will be retained for a maximum of 3 years and then appropriately deleted according to our regulations.
  • Speech recognition result information and user dictionary information set in profileWords are recorded in the operation log of the speech recognition server. These may be checked for troubleshooting and operation purposes.

Please also see Article 7 (Speech Recognition) of the Terms of Service.

About the Division of Responsibility​

To use AmiVoice API safely, both we and the customer need to implement security measures appropriate to their respective areas of management.

Our Responsibility (Service Provider)​

  • Maintenance and operation of the security of the cloud infrastructure, network, and applications included in the scope of AmiVoice API
  • Performing speech recognition processing on voice data sent to AmiVoice API
  • Maintaining security functions provided as a service provider, such as communication encryption, authentication, and access control
  • Handling of voice data, speech recognition results, and user dictionary data in accordance with the specifications of this service and the Logging option

Customer's Responsibility (Service User)​

  • Security management of the customer's systems, devices, networks, and relay systems that connect to AmiVoice API
  • Appropriate management of authentication information, including the API key
  • Verification of the content of data sent to AmiVoice API, and the rights, consent, and other legal basis necessary to send and process such data
  • Selection of the Logging option according to the content and purpose of the data being handled
  • Storage, use, deletion, and access management of the speech recognition result data obtained from AmiVoice API

For detailed regulations and related information, please also see the following pages: