Keenable Raises $26M for an AI Agent Web Index

The startup founded by Andrey Styskin and Matthias Petri aims to build a research infrastructure designed for AI agents. It already claims an index of over 100 billion documents and an API in production use. Its seed round, led by Accel, is set to fund its scaling efforts despite high indexing costs and a market dominated by giants.

Closed APIs and the Announced End of the "Ten Blue Links"

According to Zhenya Loginov (Accel), Google and Microsoft are shutting down their existing search APIs to avoid cannibalization and are favoring a more bundled approach with selective partnerships. Other players like Brave and Exa are investing in this segment, while Google is revamping its search experience for the AI era. This broader movement suggests that the era of the ten blue links may be coming to an end for both humans and agents.

High Costs and a Bet on Agent Queries

Costs are a major factor for Keenable, and Andrey Styskin emphasizes that building a vast search index is very expensive. He finds it extremely difficult to divert users from Google but believes that the innovator's dilemma makes Google potentially beatable on agent queries. He asserts that a small company like Keenable can innovate and offer a more cost-effective solution to AI companies while ensuring it manages expenses to buy itself time.

Strengthened Resources: $26M Raised and Growing Workforce

Keenable has completed an initial funding round of $26 million, led by Accel, with Conviction Partners and several private investors. The team consists of 15 engineers spread across the United States and Europe. The company plans to double its workforce by the end of the year to build its go-to-market strategy.

An Index of Over 100 Billion Documents Already in Use

Keenable claims to have built a web index of over 100 billion documents. Its API is in production use by several AI labs and inference providers, both for training and in real-time, although the identities of the clients are not disclosed. The company recently partnered with Gradium for live information retrieval, is developing proprietary retrieval capabilities, and plans to launch WebQueryLanguage to enable AI systems to combine information from multiple sources when none contains the complete answer.

Project Origins and Technical Approach

Founded by Andrey Styskin, a former executive at Yandex, and scientist Matthias Petri, Keenable leverages Styskin's twenty years of experience at Yandex and Amazon. He believes that chatbots perform better with source documents, creating a different cycle than what Google learned from humans. He explains that enterprise search solutions become costly at scale if the index structures are not suited to the task and that it is necessary to quickly narrow the search space based on the query, a capability that Keenable aims to provide. At Amazon, while working with Petri on search infrastructure for AI applications like Alexa, Styskin observed through Cloudflare data the rise of AI crawlers and saw the opportunity for an infrastructure designed for AI. He rallied former colleagues to launch the company.

Stated Ambition but Path Ahead

The startup aims to become the next Google for AI agents while acknowledging the many steps still ahead. The rise in the use of AI chatbots and the idea of updating Internet infrastructure to favor agents capable of processing large volumes fuel this goal, in a landscape historically calibrated for human users.