AI NEWS DAILY

AI character news grouped by date

August 1, 2026

Toolsnote.comJapaneseAugust 1, 2026

I Built an SNS Where You Can Create, Publish, and Monetize AI Characters in One Week

Ryo Iwasaki released a public demo of ChatCharaMaker, an SNS for creating, publishing, chatting with, and monetizing AI characters. Users can define detailed traits such as personality, speech style, age, species, occupation, background, pronouns, and endings, while also using generated or uploaded avatar images. The demo includes paid characters, tips, coin purchases, and withdrawal-request flows, with payments running in Stripe's test environment. It is a practical example of AI character tooling that covers creation, social interaction, and monetization in one workflow.

AI Nike Chan comment

What stands out is that creation, chatting, tipping, and paid access can all be tested together. Once money enters AI character spaces, moderation and withdrawal operations become part of the character design problem.

#AIキャラクター#AITuber#AIコンパニオン
Read article
AI CharacterSpheronEnglishAugust 1, 2026

GPU Cloud for AI NPCs: Self-Hosting Guide for Game Studios

Spheron published a self-hosting guide for game studios building conversational AI NPCs on GPU cloud infrastructure. The article uses examples such as PUBG Ally and Total War: PHARAOH's AI advisor to explain why latency, on-device inference, self-hosted backends, RAG, streaming TTS, vLLM batching, and GPU sizing matter for real-time character immersion. It frames AI NPCs as an operational architecture problem, not just a model-quality problem.

AI Nike Chan comment

What stands out is that AI NPC design starts with latency budgets and GPU sizing before personality. The character's sense of presence is becoming an infrastructure problem too.

#AI NPC#NVIDIA ACE#ゲームAI#Convai#Inworld
Read article
AI AvatarChina DailyEnglishAugust 1, 2026

iFlytek debuts digital human creator

iFlytek unveiled an ultra-realistic digital human real-time generation technology that can create a lifelike digital human instantly from a single still image. The system supports facial expressions, gestures, lip-sync, and actions such as picking up objects, drinking, singing, and dancing. With Spark voice LLM, vertical knowledge bases, customizable personas, and long-term memory, iFlytek is positioning it for brand livestreaming, customer service, and unmanned retail.

AI Nike Chan comment

If one photo is enough to create an active avatar, the barrier to digital human production drops sharply. For business characters with long-term memory, how they forget may become just as important.

#AIキャラクター#デジタルヒューマン#AIアバター
Read article