Keenable exits stealth with $26M seed to index the web for AI agents

A web index built for machines, not readers

Keenable, a startup founded by former Yandex search and AI executive Andrey Styskin and German AI scientist Matthias Petri, has emerged from stealth with $26 million in seed funding. According to TechCrunch, Accel led the round, with Conviction Partners and a group of business angels also participating.

The company's premise is that search infrastructure was optimized for humans, who skim pages and click links. AI agents and chatbots consume the web differently: they can process far more text and need grounded access to source documents at machine scale. Styskin told TechCrunch that grounding answers in source documents creates a new kind of flywheel, distinct from the behavioral signals Google learned from human users.

More than 100 billion documents, already in production

Keenable says it has built a search index covering more than 100 billion documents, and that its API is already running in production at several AI labs and inference providers, during both model training and runtime. The company declined to name those customers, but TechCrunch reports it recently struck a partnership with voice AI company Gradium to support live information retrieval.

The startup is also developing proprietary retrieval tooling, including an upcoming product called WebQueryLanguage, designed to let AI systems answer questions by combining information from multiple web sources even when no single page contains the full answer.

Styskin, who says he spent about 20 years building search at Yandex and Amazon, positions Keenable apart from enterprise search products that become costly and fragile at web scale. As he explained to TechCrunch, scanning the entire internet for every query is enormously expensive, so the key innovation is narrowing the search space very quickly based on the query itself.

Why the gap opened

Accel partner Zhenya Loginov, who led the investment, told TechCrunch that AI companies have very few options for web-scale search infrastructure, especially as Google and Microsoft take steps to shut down their existing search APIs to avoid cannibalizing their own businesses, favoring bundled arrangements with selected partners instead.

Styskin's conviction dates to his time at Amazon, where he worked with Petri on web search infrastructure for AI applications such as Alexa. According to TechCrunch, Cloudflare data he had seen showed AI crawlers driving a growing share of search volume, which signaled demand for search infrastructure designed with AI in mind. He then tapped his network to hire a handful of former colleagues for the new venture.

Competition and cost

Keenable is not alone in the space. Brave and Exa are also building search alternatives aimed at the AI era, and Google itself is overhauling its search experience. Styskin concedes that pulling people away from Google for ordinary search is extremely difficult, but argues the innovator's dilemma leaves the US giant vulnerable on agentic queries, where a smaller company can move faster and price more aggressively.

The economics remain the hard part. Building a massive index is, in Styskin's words, "painfully expensive," though he says the company is pacing itself and working to keep costs under control. Keenable currently fields 15 engineers across the US and Europe and plans to roughly double headcount by the end of the year to build out its go-to-market operation.

Why it matters

Keenable's raise signals a genuine infrastructure shift: as agents increasingly browse, retrieve and act on web content on users' behalf, the indexing layer underneath is being rebuilt for machine consumption rather than human attention. If Google and Microsoft keep restricting access to their search APIs, independent index providers become critical plumbing for the AI industry, and whoever controls retrieval for agents may hold the next era's equivalent of search distribution. Keenable's ambition of becoming the next Google for AI agents is far from assured, but the fact that Google is simultaneously reworking its own product suggests the era of the ten blue links is closing, for humans and agents alike.