Bridgewell Inc. released the beta version of “ToyTalk” https://toytalk.ai/, an AI platform that allows users to enjoy real-time voice conversations with AI characters, on August 8, 2025. Until now, creating an AI character in a specialized field required advanced tuning by an IT engineer. However, ToyTalk’s unique feature is that anyone can easily create an infinite number of unique characters by combining the “behavior” required by the voice AI with “specific knowledge (up to 3,000 characters).”
With ToyTalk, you can choose from a wide range of AI characters to have a conversation with, including fortune tellers, chatters, and assistants that are useful in everyday life. You can also create a historical figure (*only those whose copyrights have expired) and ask them about what things were like at the time, or have fun talking about what might have happened. You can also easily use it as a Q&A chatbot by entering the content of your Q&A in the “Unique Knowledge” field.
ToyTalk’s voice AI can converse with users as smoothly as if they were talking to a human. While it is generally said that a human responds in 0.2 seconds, ToyTalk responds and converses in just 0.3 seconds, allowing users to enjoy a conversation as if they were talking to a human, without any sense of discomfort.
Also Read: Arent Inc announces AI integration plan for software
Using AI in everyday life
Although there is a social phenomenon that could be called an AI boom, such as Chat GPT provided by OpenAI, the usage rate in Japan is lower than overseas, and even for those who do use it, it is mostly limited to work-related purposes. (*The daily usage rate of generative AI is 72% globally, but in Japan it is low at 51%. Source: “AI at Work 2025: Momentum Builds, But Gaps Remain”)
ToyTalk hopes that by making its generative AI available for free, it can be used by a wider range of people as a useful tool in everyday life. Because the generative AI is available using conversational voice, it can be used by people who are not accustomed to typing, or even when their hands are full.
Technical mechanism of AI
AI is used for a wide range of purposes, including coding, writing text, and generating images, but when it comes to AI that is specialized in a particular field, it requires long-term learning and advanced tuning, making it difficult to implement.
Asking AI to do what you want it to do requires a certain knack. It can be learned over time and with practice, but many people do not use AI because they do not know how to use it or what to do with it (*The majority of reasons for not using AI are “I don’t know how to use it” or “It’s not necessary for my life (I don’t know what to do with it)” Source: Ministry of Internal Affairs and Communications “Information and Communications White Paper”). Lowering this hurdle could lead to increased use of AI.
ToyTalk is designed so that anyone can easily give instructions (“behaviors”) to the AI character, with the AI assisting them like training wheels on a bicycle. What’s great about ToyTalk is that even people who aren’t good with IT can use it without any manuals.
SOURCE: PRTimes