Keenable, a startup founded by former Yandex search lead Andrey Styskin and AI scientist Matthias Petri, has raised $26 million to develop a vast web search index designed specifically for AI-driven applications. Their platform aims to improve AI chatbot performance by enabling better grounding of responses in web documents, addressing challenges in current internet search infrastructure.
- Keenable raised $26 million seed led by Accel for AI-focused web indexing
- Built an index of 100+ billion documents to support AI chatbots and agents
- Upcoming Web Query Language aims to combine web data for complex AI answers
What happened
Keenable has surfaced from stealth mode with $26 million in seed financing led by Accel, alongside Conviction Partners and individual investors. The startup is developing a web search index specifically tailored for AI agents and chatbots, differentiating from traditional search engines optimized for human users. Their index now reportedly covers more than 100 billion web documents.
The company’s product is already in use by several AI development labs and inference providers for both training and runtime tasks. Keenable has also established partnerships, such as with voice AI firm Gradium, to enhance live information retrieval capabilities. The startup’s team includes engineers formerly from Yandex and Amazon, bringing two decades of search expertise to bear on this challenge.
Why it matters
Current web search infrastructure is designed with human attention spans in mind, often limiting how much of a webpage is scanned and processed. AI chatbots, however, can handle far more data, creating demand for search systems built specifically to serve AI’s unique needs. Keenable addresses this gap by focusing on indexing and retrieval methods optimized for agentic use cases.
This innovation comes amid moves by major search providers like Google and Microsoft to restrict their search APIs to protect core business models, leaving AI developers with fewer options for web-scale search access. Keenable aims to provide a more cost-effective and specialized alternative, potentially challenging incumbent search engines when it comes to AI-powered queries.
What to watch next
Keenable plans to double its 15-person engineering team by the end of 2026, targeting expanded product launches and go-to-market efforts. The startup’s forthcoming Web Query Language, designed to answer complex questions by synthesizing information across multiple web sources, will be a key development to monitor.
Competition is intensifying as other companies like Brave and Exa enter the AI-focused search space, while Google is revamping its own search experience for AI integration. How Keenable differentiates itself technologically and sustains cost efficiencies will be crucial for its ambition to become a leading infrastructure provider for AI agents accessing the web.