The release date for the "Sentiment Analysis Option" of the Asynchronous HTTP Interface v2, which was scheduled to be released today, July 31, 2026, has been postponed to August 12, 2026.

During the final checks before launching the service, we identified an issue that could affect the stable provision of the service. Therefore, we have decided to fix the issue and perform operational checks before launching the service.

We sincerely apologize to our customers who have been waiting for the service to become available for the short notice.

The schedule for the future is as follows:

• Previous date: July 31, 2026
• Revised date: August 12, 2026 (scheduled)

Please note that the sentiment analysis option for the currently offered Asynchronous HTTP Interface v1 will remain available. Furthermore, other features of the Asynchronous HTTP Interface v2 will not be affected.

Important notes regarding the transition to v2

Because v1 and v2 use different versions of the sentiment analysis engine, the values ​​of the sentiment parameters returned will not be the same even when processing the same audio.

This is due to an update to the sentiment analysis engine, and there are no changes to the definition of the emotions represented by each parameter. However, if the sentiment parameters are being judged by comparing them to thresholds, the judgment results may change with the transition to v2.

If you are affected, please process your actual audio data using both v1 and v2 before migrating to v2, and verify the following points.

• How much will the values ​​of the emotion parameters you are currently using change?
• Is the currently set threshold still valid for v2?

The v1 sentiment analysis option will be available alongside v2 until December 31, 2026, the scheduled end date for the Asynchronous HTTP Interface v1.

After the release of v2, conduct comparative testing during the period when parallel use is permitted, and adjust thresholds and judgment logic as needed before migrating.

We sincerely apologize again for any inconvenience and trouble this may cause our customers.

We appreciate your continued support of the AmiVoice API.