Livetoon's Tatara Turns User 3D Characters into Trainable AI
Tatara shifts AI characters from company-prepared selections to user-created ones, opening the technology Livetoon built for kaiwa so creators can train characters themselves.
Reporting from 1 source: ASCII.jp.
Livetoon announced Tatara, a service that lets users set a voice, personality, tone, and role for their own 3D character and raise it into an AI character capable of conversation and tasks. The service is currently invitation-based, with invitations sent sequentially to applicants. Official release is scheduled for September 1, 2026. A free early access campaign on the official X account starts August 1, 2026. Tatara builds on technology from Livetoon's kaiwa voice conversation service.
Tatara starts with a model the user owns. Users load a VRM file they hold usage rights to, then set its voice, personality, way of speaking, name, and role. Livetoon calls that process "training a character."
On desktop, the character answers by voice, with facial expressions and motions synced to the words. Through AI tool integration, users can verbally request tasks such as coding and check progress in conversation; the initial version supports OpenAI Codex.
The underlying tech comes from kaiwa, Livetoon's voice conversation service with 3D AI characters. With kaiwa the company made the characters. With Tatara it opens that system to creators.
Initial verification members get the service free for a period, and their feedback feeds the next development phase. The early access campaign on the official X account begins August 1, 2026.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.