Search on the web was designed around human habits: people type a query, scan a page of results, open a few links, and decide what looks useful. Keenable is betting that AI agents need something different.
The startup, founded by Andrey Styskin and Matthias Petri, has come out of stealth with $26 million in seed funding. Accel led the round, with participation from Conviction Partners and some business angels.
A Search Index Built For Machines
Keenable is working from a simple premise: AI chatbots and agents can process much more information than people, but they still need reliable source material. Styskin told TechCrunch that AI chatbots tend to perform better when their answers are grounded in source documents.
That changes the job of search infrastructure. A search engine for a person is optimized to help a reader choose from visible results. A search system for an AI agent must help software retrieve, compare, and use information at a much larger scale.
Keenable says it has built a web search index of more than 100 billion documents. The company also says its API is already used in production by several AI labs and inference providers during both training and runtime, although it has not named those customers.
One named relationship is Gradium. Keenable recently formed a partnership with the voice AI company to support live information retrieval, a use case where fresh and accessible web information can matter during an interaction.
Why Existing Search Infrastructure Is Under Pressure
The company is entering the market at a moment when the role of search is being reconsidered. People are increasingly using AI chatbots to look up information and complete tasks. That creates demand for infrastructure that can serve AI systems directly, rather than only presenting links to people.
According to Accel partner Zhenya Loginov, AI companies have limited choices for web-scale search infrastructure. He pointed to Google and Microsoft taking steps to shut down their existing search APIs to avoid cannibalization, while moving toward more bundled approaches and selective partnerships.
For startups building AI products, that matters because search is not only a user interface. It is also a supply chain for information. If access to web-scale retrieval is limited or bundled into broader platforms, AI companies may need alternatives that are built for their own workflows.
Styskin saw the opening while working at Amazon with Petri on web search infrastructure for AI applications such as Alexa. He had also seen Cloudflare data showing that AI crawlers were responsible for a growing share of search volume. From there, he concluded that AI-first search infrastructure could become its own category.
The Cost Problem Keenable Wants To Solve
Building a search index at internet scale is expensive. Styskin was direct about that challenge, saying, “Don’t ask — it is painfully expensive.”
The company’s argument is that cost cannot be solved only by adding more infrastructure. Styskin said web-scale systems need index structures tuned for specific tasks, because scanning and serving the whole internet without narrowing the search space quickly becomes enormous in cost.
That is where Keenable wants to differentiate. The startup says it is building proprietary retrieval capabilities that can quickly reduce the search space based on a query. In plain terms, the goal is to help AI systems find the relevant part of the web without paying the full cost of looking everywhere every time.
Styskin also sees a business opening in what he described as the innovators’ dilemma. He said it is “extremely hard” to convince people to move away from Google for search, but he believes Google may be “beatable” on agentic queries if a smaller company can move faster and offer a more cost-efficient product for AI companies.
What Comes Next
Keenable is not only offering an index and API. It is also developing an upcoming product called Web Query Language, which is meant to help AI systems answer questions by combining information from multiple web sources when no single source contains the complete answer.
The company currently has a team of 15 engineering staff across the U.S. and Europe. It plans to use the new funding to double its headcount by the end of the year and build its go-to-market motion.
The competitive landscape is already forming. Brave and Exa are also active in the space, and Google itself is changing its search experience for the AI era. Keenable’s ambition is large: becoming “the next Google for AI agents.”
Whether that happens is unresolved. What is clearer is that the web is being reorganized around a new type of user. Humans still search, but agents are becoming search users too. If that shift continues, the old model of results pages and the era of the “ten blue links” may give way to infrastructure that is invisible to people but central to how AI systems understand the web.