Keenable Exits Stealth With $26M Seed Round to Index Web for AI Agents

Keenable has officially exited stealth mode, securing $26 million in seed funding to build a web search index designed specifically for AI agents. The company aims to solve the efficiency bottleneck facing AI models that require real-time grounding in global web data.

The funding round was led by Accel, with additional participation from Conviction Partners and various business angels. Co-founded by Matthias Petri, a German AI scientist, and Andrey Styskin, the former head of Yandex's search, AI, and cloud division, Keenable has already constructed a massive index containing more than 100 billion documents. Unlike traditional search engines, this infrastructure is optimized for the processing patterns of chatbots and AI labs rather than human scanning habits. The company has already established a partnership with voice AI firm Gradium to facilitate live information retrieval.

The Infrastructure Gap

Traditional search engines are built for humans who click through a list of links, but AI agents can ingest and process far larger volumes of data simultaneously. This creates a demand for a different kind of retrieval infrastructure—one that allows bots to ground their responses in source documents more efficiently and at a lower cost.

This market opportunity is further widened by the behavior of industry incumbents. Tech giants like Google and Microsoft have increasingly restricted their search APIs to prevent third-party AI tools from cannibalizing their own bundled AI offerings. By providing an open, agent-centric alternative, Keenable is positioning itself as a critical utility for the AI ecosystem.

Why It Matters

Keenable is effectively attempting to become the "Google for AI agents," targeting the high cost and technical friction of web-scale retrieval. The financial burden of scanning the internet is a primary hurdle for AI developers. Andrey Styskin noted that without fine-tuning index structures for specific tasks, the cost of serving and scanning the entire internet is "enormous" due to the sheer volume of data. When asked about the specific costs of building such a giant index, Styskin described the expense as "painfully expensive."

By innovating on index structures specifically for agentic queries, Keenable provides a scalable alternative to the restrictive APIs of the major tech platforms. This could fundamentally shift how AI models access and synthesize real-time information, moving away from closed ecosystems toward specialized retrieval layers.

What's Next

As AI labs and inference providers begin integrating Keenable's API for training and runtime grounding, the industry will be watching to see if a specialized index can truly outperform general-purpose search for agentic workflows. While the company has secured its initial funding and a massive data set, the long-term challenge will be maintaining the freshness and accuracy of a 100-billion-document index in an ever-evolving web environment.