Keenable exits stealth mode with $26M seed round to build search infrastructure for AI agents

AI search startup Keenable has raised $26 million in seed funding led by Accel as it launches search infrastructure designed to give AI agents cheaper and more continuous access to information on the web.

Conviction and angel investors from companies including Amazon, ClickHouse, Databricks, Google, Snowflake and SpaceXAI also participated.

According to Keenable, traditional search APIs are increasingly unsuitable for AI because they were designed for humans making occasional queries rather than agents that may repeatedly search the web as part of complex tasks. Those searches can quickly become expensive, encouraging AI models to rely on training data instead of retrieving current or specialized information.

Keenable was founded by former Yandex and Amazon AGI search engineers Andrey Styskin and Matthias Petri. Styskin previously led Yandex Search and built a 200-billion-document index, while both later worked on large-scale web-search infrastructure at Amazon.

The startup operates an independent index of more than 100 billion documents and is developing retrieval technology specifically for AI workloads. Its Web Query Language allows models to search and connect information from thousands of sources, potentially helping them answer questions that require synthesizing information across the web.

The company says it has signed commercial contracts with several AI labs and is working with model builders and inference providers on runtime and training applications.

The new funding will support its engineering team, web-crawling infrastructure, index and retrieval technology.