On Monday, Eugenia Kuyda opened Wabi to all users: the company’s new messenger-style personal AI lets people maintain multiple agentic threads dedicated to different tasks and long-term goals, and run custom, built-in mini apps inside each thread. Examples include a weight-loss tracker in a health thread or an app listing nearby children’s events in a family thread.
Roots: Luka and the Replika legacy
Kuyda’s startup began as Luka. Twelve years ago she demonstrated a conversational chatbot called IO at TechCrunch Disrupt that acted as a lifestyle concierge to help with bookings and travel. Luka later evolved into Replika. Kuyda left her CEO role at Replika last year and founded Wabi.
Today’s advances in deep learning make the original Luka vision more feasible: modern language models allow Wabi to offer capabilities that earlier chatbots could only hint at.
A pivot or an evolution: how Wabi changed
Wabi’s public launch follows a period of product evolution. When TechCrunch covered Wabi’s initial launch last year and its $20 million pre-seed round, the product was described more like a social platform where users could create and share mini apps composed of prompts.
Over the past year, the team concluded that while people prefer personal software to one-size-fits-all App Store solutions, many do not want to spend the time building it themselves. In Wabi 2.0, the agent decides which apps to build: once you connect your email, it automatically spins up several starter threads and apps, and users can request custom apps or open new threads.
Kuyda says the approach overlaps with what competitors like Muse and Instinct promise, but the combination of built-in apps and multi-thread organization is a key differentiator.
Multi-threaded and multiplayer design
Wabi’s notable design choice is its multi-threaded, multiplayer architecture. Kuyda argues that single-thread experiences — a continuous, ever-growing conversation with one agent as seen in some rivals — degrade over time because too much context accumulates in one thread, increasing the risk of hallucinations, context loss, or unsafe responses.
Wabi’s threads can be shared and function like group chats. Kuyda says this encourages collaborative use rather than everyone separately interacting with their own AI. In her demonstrations she showed a thread she shares with her boyfriend where a row of apps at the top helps them coordinate activities: a live music tracker, a daily-relationship-question app, a running list of restaurant suggestions, important shared calendar events, and a shared reading list with recommendations.
In a family-focused thread she showed apps tracking her daughter’s schedule, surfacing nearby kid-friendly events, and a Japanese quiz for practice. Within any shared chat, participants can choose how much personal data to share — from email and calendar information to bank account details — and access is scoped to the individual thread. Everyone in the chat can see when someone requests another person’s personal information.
Privacy, operation and business model
Wabi is SOC 2 compliant, meaning its data security controls have been independently audited. The service runs entirely in the cloud, and most user data is stored in the U.S., according to Kuyda.
Wabi is free for now, and the company has not settled on a final business model. Kuyda said Wabi is in talks to raise more funds.
Why this matters
Wabi’s launch marks a notable entry in the personal-AI space where startups compete to automate daily organization, improve efficiency, and support shared social experiences. Its multi-thread and multiplayer approach and the claim that the agent will build starter apps seek to address two common frictions: users’ lack of time to construct custom personal software, and the problems that arise in long, single-thread conversations with chatbots.
Kuyda’s prior work with Replika focused on building AI that helps reduce loneliness; with Wabi she aims in part to use AI to help people spend higher-quality time together in real life, whether finding concerts, coordinating family schedules, or choosing books to read together.
