Twelve years after Eugenia Kuyda stood on a TechCrunch Disrupt stage and demoed IO, a conversational lifestyle concierge chatbot, the technology has finally caught up with her original vision. IO was the first product from Kuyda’s startup Luka, built before the deep learning breakthroughs that produced today’s large language models. Kuyda told TechCrunch that IO still worked and was ‘pretty magical,’ but far from what exists now. Luka eventually became Replika, the AI companion company Kuyda left as CEO last year to found Wabi. On Monday, Wabi launched for all users as a multiplayer personal AI assistant built as a messenger app: users maintain multiple agentic threads dedicated to different tasks and long-term goals, and within each thread they can build or request custom apps, such as a weight-loss tracker in a health thread or an app showing nearby children’s events in a family thread.

The launch is something of a pivot; Kuyda frames it as an evolution. When TechCrunch covered Wabi‘s initial launch last year, alongside a $20 million pre-seed round, it looked like a social platform where people could create mini apps with prompts and share them with friends or across the app. Over the following year, the team realized people love personal software over one-size-fits-all App Store solutions, but ‘maybe don’t want to spend a lot of time, or any time, building it.’ So in Wabi 2.0, the agent decides what apps to build. The moment a user connects email, Wabi automatically spins up a number of threads and apps to start with, and users can add more by requesting custom apps or starting new threads. According to Kuyda, this means Wabi does what assistants like Instinct and Muse do, but with an app layer that helps users do more.

Kuyda is also betting on multi-threaded, multiplayer design. She argues that single-thread experiences get worse over time because one endless conversation accumulates too much context. AI researchers interviewed for this story agree that longer chatbot conversations are more likely to produce hallucination, context loss, or responses blocked by safety guardrails. Wabi threads can be shared and function as group chats, letting multiple people interact with the same agent together. Kuyda said that today all agentic apps are just conversations between one user and an agent, but the real opportunity is bringing people together in a group chat—’I don’t think anyone has yet cracked multi-player.’ She demonstrated with her own shared threads: in one with her boyfriend, apps along the top include a live music tracker, a daily relationship question, a running list of restaurant suggestions, important events from their shared calendar, and a shared reading list with next-book suggestions. In a family-focused thread, she has an app tracking her daughter’s schedule, another finding kid-friendly events nearby, and another that quizzes her daughter on Japanese.

Access to personal information is scoped to individual chats. Users decide how much to share—from email and calendar to bank account details—and everyone in the chat can see when someone’s personal information is requested. Her partner can ask Wabi directly ‘Is Eugenia free next Thursday?’ or ‘Did she buy those tickets?’ On the infrastructure side, Wabi is SOC 2 compliant, runs entirely in the cloud, and most user data is stored in the U.S. Kuyda connects the product to Replika‘s earlier mission of using AI to reduce loneliness: Wabi can bring people together by helping them spend time better in real life, whether that is going to a concert together or choosing the next book to read together. Wabi is free for now, the company hasn’t settled on a business model, and Kuyda says Wabi is in talks to raise more funds.