Now exiting stealth mode with a $26 million seed round, Keenable has been building a vast web search index for AI agents.

Backed Keenable builds web search infrastructure for AI agents. Discover more Web Apps & Online Tools Business Operations Space Technology Accel-Backed Keenable builds web search infrastructure for AI agents. Keenable raises $26 million from Accel to build web search infrastructure and indexing technology designed for AI agents. Aug 25, 2026 - 13:31 Keenable, a startup founded by former Yandex search executive Andrey Styskin and AI scientist Matthias Petri, has emerged from stealth with $26 million in seed funding to build web search infrastructure designed for artificial intelligence agents. Small Business Discover more Business & Industrial Machine Learning & Artificial Intelligence Business Formation The company is developing a large-scale web index that allows AI systems to retrieve information from online sources more efficiently. Keenable said its technology is designed for AI applications that require access to up-to-date information and source documents during training and in real time. Keenable builds search infrastructure for AI systems. Traditional search engines were designed around human users who typically scan search results and individual web pages. However, as AI assistants and agents increasingly perform web-based tasks, Keenable believes the internet infrastructure needs to adapt to support systems capable of processing much larger volumes of information. Styskin, who previously led search, AI, and cloud operations at Yandex, said AI systems often perform better when source documents support responses. He founded Keenable with Petri to develop search technology designed specifically for AI-driven applications. The startup said it has created a web search index containing more than 100 billion documents. Several AI labs and inference providers already use its API during training and runtime, although Keenable has not publicly identified those customers. Small Business Building retrieval tools for AI agents. Keenable recently partnered with voice AI company Gradium to support live information retrieval during conversations. The company is also developing proprietary retrieval technology, including an upcoming product called Web Query Language that would allow AI systems to combine information from multiple web sources. Discover more Small Business Styskin said Keenable's approach differs from traditional enterprise search solutions because serving search results at internet scale requires specialised indexing methods that can quickly narrow the search space based on a query. Existing web search access options have also been changing. Changes to Google's Custom Search API and Microsoft's Bing Search API reflect broader shifts in how major search providers provide access to their infrastructure. Accel partner Zhenya Loginov, who led the investment, said AI companies have limited choices for web-scale search infrastructure as large technology companies increasingly move toward more selective and bundled approaches. Keenable targets search built for AI workloads. The founders said their interest in AI-focused search infrastructure stemmed from prior work building search systems for AI applications, including voice assistants. Styskin said he began seeing an opportunity as AI crawlers became a larger part of web activity. Cloudflare's analysis of AI crawlers highlighted the growing prevalence of AI-related web crawling, adding to interest in infrastructure designed for machine-based information retrieval. Styskin said the startup is also exploring opportunities arising from what he described as the "innovator's dilemma," in which established companies may struggle to adapt existing systems to new markets. The concept was outlined in The Innovator's Dilemma by Clayton Christensen. Small Business Keenable is entering a competitive market that includes other AI search companies and evolving search products from established providers. Companies such as Brave Search and Exa are also developing search technologies aimed at AI applications. Startup expands team after $26 million funding. Building and maintaining a large web index requires significant resources. Styskin acknowledged that creating this infrastructure is expensive, but said the company is focused on managing costs while expanding its technology. Keenable currently has a team of 15 engineers across the United States and Europe. The startup plans to use its new funding to double its headcount by the end of the year and build its go-to-market operations. The company's long-term goal is to create search infrastructure designed for AI agents rather than traditional web users. However, Keenable faces competition from both emerging AI search companies and established search providers adapting their products for the AI era.

Keenable raises $26M seed to index the web for AI. Published · Aug 25, 2026 Why Relve is watching this Dedicated web-scale retrieval for agents is emerging as a distinct layer as incumbents bundle search. Over the next quarters, customer traction and cost curves will reveal whether standalone indexes can compete. Key Takeaways * Keenable raises $26M seed led by Accel to build a web index tailored for AI agents, exiting stealth. * 100 billion documents are indexed, and the API is used in production by AI labs and inference providers during training and runtime. * Gradium partnership adds live information retrieval for voice AI use cases. * WebQueryLanguage is in development to combine answers across multiple web sources. * Scarce alternatives at web scale, as Google and Microsoft restrict search APIs, underpin Accel's thesis. What happened. Keenable raises $26M seed to index the web for AI agents, exiting stealth with a bid to supply retrieval infrastructure that grounds chatbot answers in source documents. Founded by Andrey Styskin and German AI scientist Matthias Petri, the round was led by Accel with Conviction Partners and angels participating, TechCrunch reported. The startup says its web index spans more than 100 billion documents, with an API already used in production by several AI labs and inference providers during training and runtime. Keenable is also partnering with voice AI firm Gradium, and is developing a WebQueryLanguage to combine answers across sources. Accel's Zhenya Loginov said options are scarce as Google and Microsoft restrict search APIs. Why it matters. If AI agents are to replace scanning pages, they need fast, cheap web-scale retrieval built for machines, not humans. That aligns with a broader shift toward agent-centric interfaces, and grounding can cut hallucinations that derail decisions, with retrieval already used during training and runtime. But building and serving a massive web index is costly, and incumbents are hardly standing still as Google overhauls search and challengers Brave and Exa push alternatives. With customers undisclosed, Keenable still has to prove economics and go-to-market durability. "This actually creates a new flywheel that is different from what Google learned from human behavior." Andrey Styskin, Keenable Bottom line. The near-term markers are customer disclosures, pricing around API usage, and whether WebQueryLanguage ships and gains adoption as builders test agentic retrieval against bundled offerings. Addressing engineering teams' difficulty verifying AI outputs will be a credibility hurdle. Founders and engineering leads will weigh third-party web indexes against building in house, balancing quality, latency and cost. Vendor evaluations will intensify, and per Relve, an AI tools intelligence platform, teams can track fit and cost performance as infrastructure separates from apps.