Eugenia Kuyda is a technology founder best known for creating Replika, one of the earliest mainstream AI companion apps, and later founding Wabi, an AI platform built around personal software and agents.
Her work has moved through several generations of consumer AI: restaurant chatbots, emotional companions, prompt-built mini apps and now agents that can build an interface around the task a user is trying to complete.
What is Eugenia Kuyda known for?
Kuyda became widely known through Replika.
The product was built around the idea that an AI could be more than a search tool or productivity bot. It could remember a person, hold an ongoing conversation and become part of a user’s emotional life.
That approach made Replika influential and controversial at the same time.
How did Replika start?
Replika’s own origin story traces the project to Kuyda’s experience after the death of her close friend Roman Mazurenko in 2015.
She used a collection of his old messages to build a chatbot that could mimic parts of his conversational style.
The experiment eventually grew into a broader product where users could create their own AI companion rather than talk to a simulation of one specific person.
What did she build before Replika?
Before Replika, Kuyda co-founded Luka.
Luka initially built conversational products around practical tasks, including a restaurant recommendation chatbot.
The company’s work gradually shifted toward the more personal conversational experience that became Replika.
How big did Replika become?
Replika grew to tens of millions of registered users and became one of the most recognisable AI companion brands before the current generative-AI boom.
It also became an important case study in what happens when users build strong emotional relationships with software.
Why was Replika controversial?
The deeper an AI relationship feels, the more sensitive questions become around safety, privacy, dependency and product changes.
Replika faced regulatory scrutiny in Europe and criticism when changes to its behaviour altered relationships users felt they had already built with their companions.
Those episodes helped turn consumer AI companionship into a serious policy topic rather than only a novelty.
What is Wabi?
After Replika, Kuyda moved into a different consumer-AI idea with Wabi.
Wabi originally let users describe small apps they wanted and have AI generate them.
The examples were intentionally personal: a weekend planner for a group chat, a contraction timer, games for a child or a tiny tool built for one household rather than millions of users.
What changed with Wabi 2.0?
In September 2026, Wabi shifted toward what Kuyda calls a personal agent and messenger.
The new idea is that a user chats naturally, and Wabi creates whatever interface the situation needs while also helping complete the task.
Instead of opening a separate app builder first, the software can build a mini app inside the conversation.
Why does Kuyda call this personal software?
Traditional software is built once and sold to a large market.
Kuyda’s Wabi thesis is that AI can make software cheap enough to create for one person, one family or one temporary situation.
A tiny app does not need millions of downloads to be useful if it can be generated in seconds for the person who needs it.
How is Wabi different from Replika?
Replika is centred on the relationship between a person and an AI companion.
Wabi is centred more on action and creation. The agent can help organise something, perform a task or generate a custom interface.
Both products still share Kuyda’s interest in making AI feel personal rather than purely technical.
Why is she important in consumer AI?
Kuyda was building persistent conversational AI years before ChatGPT made chat interfaces mainstream.
Her career is useful because it shows how consumer AI has changed: first conversation itself was the product, then the industry moved toward assistants, and now many companies are racing toward agents that can take actions.
What is Wabi trying to become?
Kuyda has described the long-term goal as an operating system for the agentic era where software adapts around the person instead of forcing the person into a fixed app.
Whether users actually want thousands of temporary mini apps remains an open question, but the idea is now part of a wider industry shift toward more personalised software.
For our full product explainer, see Wabi 2.0 when it publishes later this month. For more founder profiles, see Holyprofweb’s People section.