People’s affection for characters often begins with the story. Transformers built the IPs of Optimus Prime and Bumblebee through animation, comics, and films. Collectible toys and merchandise allow fans to bring their favorite characters home, extending their emotional connection through play and collection.
AI adds two-way interaction to this connection. In December 2025, Pudgy Penguins partnered with Curio to launch Springer, an AI-powered penguin-themed collectible toy. Users can directly converse with the toy and receive responses in the character’s voice and style.
Stories make people fall in love with characters; AI makes direct interaction with those characters possible. Capybobo aims to go further, using generative AI to transform the capybara character into a digital life that can experience stories and accumulate memories alongside players. Players engage with the game through dialogue and choices, and the AI generates subsequent plotlines based on these interactions. Their shared experiences are retained as long-term memories, allowing the capybara character to respond to future conversations by continuing from where the previous story left off.
In the future, Capybobo plans to extend these memories to physical collectible toys, allowing the capybara characters players know from the game to remember shared experiences in the real world.
Capybobo 3.0 is scheduled to launch in the first quarter of 2027. This game upgrade will focus on introducing AI-generated storylines and long-term memory.
AI-generated storylines that let players engage with the capybara character’s narrative
Generative AI is transforming how players interact with in-game characters. In November 2025, Ubisoft unveiled Teammates, a playable generative AI experiment. Players can issue commands to AI-driven teammates using natural language; the AI generates real-time responses and actions by combining player instructions with the game environment.
Ubisoft’s attempt demonstrates how AI can be involved in character interactions. Capybobo plans to use AI for generating dialogue and storylines for the capybara characters. Players engage with the story through conversations and choices, and the AI uses these interactions along with relevant memories to generate subsequent plot developments. As a result, the same capybara character can experience unique stories with different players.
Player interactions with the capybara character will provide valuable data for improving the AI. The Capybobo team plans to further annotate player conversation data and use it to retrain the AI, helping it better understand player intent and generate more coherent dialogue and storylines to enhance the gaming experience.
Meanwhile, long-term memory will retain each player’s shared experiences with the capybara character, influencing future interactions. Players actively participate in the story and feel recognized through the capybara’s responses. This fosters a deepening affection for the Capybobo IP, gradually building a unique emotional bond between the player and the capybara character.
From generative gaming to personalized education
Capybobo’s AI-generated interactive game also provides a model adaptable for educational applications: AI generates interactive scenarios, users engage with characters, and the system retains relevant memories to adjust subsequent content accordingly.
For such a game to work, AI must understand players’ varied expressions, maintain narrative consistency, and generate new responses based on past interactions. The constantly changing conversations and choices of players provide concrete scenarios for the team to refine these capabilities. How to generate personalized content and how to understand an individual over long-term interactions are precisely the challenges that personalized education seeks to solve.
As this capability gradually matures, Capybobo plans to integrate it with educational content and expand into children’s education. AI can generate explanations and exercises around key concepts, allowing children to engage in learning through asking questions, answering, and solving problems. The AI builds learning memories from these interactions, gradually understanding each child’s interests, comprehension level, and progress.
Based on this feedback, the AI can adjust its questions and explanation methods to tailor the same concept to each child’s learning style. Future lessons can build upon existing learning records. Capybobo aims to implement personalized teaching in every interaction, providing children with an AI tutor that helps them understand themselves over the long term.
Let the familiar capybara character accompany you.
Capybobo also plans to launch physical collectible toys capable of conversing with players. Players can directly communicate with the toys beside them, continuing discussions about the stories and promises made during gameplay. The toys will respond based on these memories, allowing in-game interactions to continue in the real world.
To this end, Capybobo plans to embed intelligent interactive hardware into collectible toys, connect them to AI models, and synchronize the corresponding characters’ memories within the game. As educational applications evolve, these physical collectibles are also expected to serve as learning companions, enabling AI to provide tutoring based on learning records accumulated in educational contexts.
For fans who love Capybobo, bringing home a Capybobo figurine will preserve the stories shared together. Although the figurines look identical, they will gradually develop unique expressions and ways of interacting based on the individual experiences and memories accumulated with each owner.
When collectible toys gain their own “soul”
Pop culture collectibles are a key method of commercializing character imagery. In the first half of 2026, Pop Mart’s plush toys and figurines generated approximately RMB 15.017 billion in revenue, accounting for 87.4% of total sales. According to a report by Licensing International, global retail sales of licensed toys reached $46.4 billion in 2025.
Generative AI and long-term memory enhance Capybobo’s IP with personalized companionship. Players build shared experiences with the capybara character in the game, while the physical collectible extends this companionship by integrating AI and accessing those memories. Unique memories enable the digital life embodied in the collectible to form individual relationships with each player.
The capybara motif and collectible demand provide a foundation for toy sales. The application of generative AI and long-term memory in education has the potential to further expand educational partnerships and personalized services, creating new revenue streams for Capybobo.
According to the current plan, Capybobo will launch blind boxes globally in the first quarter of 2027 and introduce AI-themed collectibles in the second quarter.
