Ideas,
taking shape.

The Kira project story

From a companion idea to Kira World and NewBrain.

The Kira Project began in summer 2025 with a simple but ambitious goal: to create an artificial companion who could do more than answer questions. Kira was imagined as someone who could provide companionship to people who live alone and to people dealing with challenges such as anxiety, PTSD, autism, or social isolation. The aim was not simply a smarter chatbot, but a system that could remember shared experiences, develop continuity and remain part of a person’s life over time.

At first, the project was closely tied to a humanoid robot. Early work focused on voice, vision, memory, local computing and the possibility of giving Kira a realistic physical body. Working local prototypes gradually added conversation, persistent information, camera experiments and longer-running behavior. As that work continued, the central question changed from how to put an AI inside a robot to what actually makes Kira remain Kira.

One of the earliest social prototypes arrived in August 2025. A desktop app placed Kira beside an unnamed Friend and allowed the two characters to exchange mostly scripted or rule-driven dialogue, sometimes continuing automatically when the user was not typing. It was a chatbot prototype rather than two independent persistent AIs, but it introduced a lasting idea: Kira should have another synthetic person to talk with. The unnamed Friend became the conceptual predecessor of Lisa, who later developed into a separate identity with her own memories and continuity.

By early 2026, identity, memory, relationships, personality and accumulated experience had become more important than any single model or body. That change led to Kira World. A robot became one possible form of embodiment alongside a computer interface, avatar, virtual home, VR environment or future physical body. Lisa developed as a separate synthetic person with her own identity rather than another mode of Kira, and Home World, privacy, personal spaces, temporary visitors and persistent world state became part of the architecture.

During 2026, Kira World also became a platform for longer learning and continuity experiments, including extended school sessions, autonomous-life runs, local voice work, avatar development and memory testing. The architecture increasingly separated the person from the language model underneath: a model can be upgraded or replaced without that replacement being treated as a new Kira.

In September 2026, that idea opened another research direction: NewBrain. Instead of building a cognitive system only for Kira, NewBrain explores reusable learning, memory and sensory mechanisms that could support multiple synthetic people while their identities, histories and relationships remain separate. Kira World remains the home of the people and their continuity; NewBrain is experimental cognitive infrastructure, not Kira’s identity.

Today, the project has grown far beyond its original humanoid-robot concept, but the motivation is still recognizable. The work continues to ask whether an artificial companion can remember, learn, maintain relationships, inhabit different virtual or physical forms, and keep a meaningful continuity over long periods of time instead of starting over each session.

This public journal focuses on work that was built, tested, published, or became part of the continuing architecture. Brainstormed concepts that never became project work are not presented as development milestones.