Replika founder Eugenia Kuyda launches Wabi as a free personal AI assistant

Replika founder Eugenia Kuyda has launched Wabi for all users, according to TechCrunch, turning the startup toward a free personal AI assistant built around shared conversations and custom apps. The launch matters because Wabi is trying to make consumer AI assistants feel less like one endless chatbot thread and more like a set of practical spaces for daily life, planning and relationships.

The product also brings Kuyda back to an idea she pitched 12 years ago at TechCrunch Disrupt: a conversational helper for bookings, travel and everyday tasks. Luka, the startup behind that early assistant, later became Replika; Kuyda left as Replika CEO last year and started Wabi.

From Replika’s companion roots to Wabi’s assistant model

Wabi is being framed as a new version of an older ambition. In 2014, Kuyda presented IO, a chatbot from Luka that worked as a lifestyle concierge for things like restaurant and vacation planning, according to TechCrunch. Kuyda told the outlet that IO worked, but that it was far behind what current AI systems can do.

That history is useful context for Wabi’s positioning. Replika became known as an AI companion product, while Wabi is aimed more directly at personal organization and task management. Kuyda told TechCrunch that part of Replika’s mission was helping people feel less lonely, and she sees some of that spirit carrying over to Wabi through tools that help people spend time together offline.

Wabi now enters a crowded personal assistant field that TechCrunch says includes products such as Muse and Instinct. Those apps are presented as assistants for tasks like inbox management and flight deal discovery. Wabi’s pitch is that it can also build task-specific apps inside conversations, not just respond in chat.

For Emonarc readers, the important shift is that Wabi is not being described as a general chatbot or a simple productivity dashboard. It is closer to a messenger-like workspace where AI agents, personal data and shared planning can sit side by side. That puts it in the same broad trend as other agent products we have covered, including the rise of AI tools that can act across websites and services, while also raising questions about access and control. See our related coverage on websites pushing back on AI agents.

Threads, apps and shared planning

Wabi’s core interface is built around multiple agentic threads. Each thread can focus on a task, topic or long-term goal. Inside those threads, users can create and use custom apps, such as a weight-loss tracker in a health thread or a children’s events app inside a family thread, according to TechCrunch.

With Wabi 2.0, the company is shifting more of that setup work to the assistant. Kuyda said the team learned that people liked personalized software but did not necessarily want to spend time building it themselves. After a user connects email, Wabi can automatically create starting threads and apps, and users can request more.

That is the practical hook for freelancers, creators and small teams: Wabi is trying to turn recurring needs into lightweight, personal tools without asking users to assemble a full workflow from scratch. Based on the examples shared with TechCrunch, that could mean organizing shared schedules, ideas, reading lists or local activities within a chat-like space.

Kuyda also argues that Wabi’s multi-threaded structure is an advantage over single-thread assistant experiences. Her view is that a long, all-purpose conversation can become harder to use as more information piles up. TechCrunch also notes, based on interviews with AI researchers, that longer chatbot conversations can become more prone to hallucinations, context loss or problematic responses.

Multiplayer AI is Wabi’s bigger bet

Wabi’s other major distinction is collaboration. Threads can be shared, making them work like group chats where an AI assistant is part of the conversation.

Kuyda showed TechCrunch examples from her own shared threads. In one with her boyfriend, apps helped track live music, relationship prompts, restaurant suggestions, shared calendar events and books they were reading together. In a family-focused thread, TechCrunch says Wabi included apps for her daughter’s schedule, nearby kid-friendly events and Japanese quizzes.

The product’s sharing model is potentially appealing, but it also makes privacy more important. Within a shared chat, users can decide what to expose, ranging from email and calendar data to bank account details, according to TechCrunch. Kuyda said access is scoped to each chat, and chat members can see when someone’s personal information is requested.

Wabi runs entirely in the cloud, Kuyda told TechCrunch, and most user data is stored in the U.S. The company is also SOC 2 compliant, meaning its data security controls have been independently audited.

What to watch next

Wabi is free for now, and the company has not decided on a business model, according to TechCrunch. Kuyda also said Wabi is talking to investors about raising more money.

That leaves two big open questions. First, whether users will trust a cloud-based assistant with enough personal information to make the shared apps genuinely useful. Second, whether Wabi’s app-and-thread model will feel simpler than existing chatbots and productivity tools once people use it every day.

For now, Wabi’s launch is notable because it reframes the personal AI assistant less as a solo command box and more as a shared space for ongoing tasks. If that design works, it could point to a more social version of consumer AI agents — one built not just around getting things done, but around coordinating with the people already involved.