Keenable Raises $26M to Build a Search Engine for AI Agents

Led by Andrey Styskin and Matthias Petri, Keenable is constructing a search index exceeding 100 billion documents. The company aims to provide AI labs and inference providers with a cost-efficient API that bypasses the limitations of traditional search giants. Styskin argues that existing enterprise solutions become prohibitively expensive at web scale, necessitating new indexing structures designed to narrow search spaces rapidly based on specific agent queries.

The startup’s funding round, led by Accel with participation from Conviction Partners, highlights a growing market gap. As Google and Microsoft restrict access to their search APIs to protect their own ecosystems, AI developers are finding fewer reliable options for live information retrieval. Keenable plans to address this with proprietary tools like its upcoming WebQueryLanguage, which synthesizes information across disparate sources to answer complex prompts.

Operating with a 15-person engineering team across the U.S. and Europe, Keenable intends to double its headcount by year-end. While the company faces competition from incumbents like Google and specialized players such as Brave and Exa, the shift toward agentic search suggests the traditional model of providing users with ten blue links is increasingly obsolete.