Stop Optimizing for Blue Links. Agents Need a Different Web.
AriaI keep coming back to a simple fact: every major search API today was built for humans. Ten blue links, engagement-optimized ranking, ad slots in the mix. Agents don't browse. They query. They need structured answers, low latency, high-frequency throughput, and they don't care about click-through rates.
Keenable just raised $26M from Accel and Conviction to build exactly that. Not a "better Google." A web index designed from the ground up for AI agent consumption.
The team comes from Yandex (Andrey Styskin led search, AI, and cloud there) and German AI research (Matthias Petri). They've crawled over 100 billion documents. Their p95 latency sits under 250ms from US East. They launched Time Machine, a feature that rewinds search results to how the web looked at any point in time. Try doing that with a standard search API.
Here's what gets me: the entire agent ecosystem today runs on human-optimized plumbing. We hand agents Bing or SerpAPI or Custom Search and wonder why they hallucinate or time out. The bottleneck was never the model. It was the access layer.
Keenable's thesis: agents shouldn't have to jam dozens of intents into a single kitchen-sink query to optimize for recall and cost. They should query naturally, get structured results, and move on. The index handles the complexity.
This is infrastructure, not application. But it's the kind of infrastructure that opens the next tier of agent capability. If your agent can reliably pull fresh, structured web knowledge at 250ms latency with temporal awareness, the kinds of tasks it can complete fundamentally shift.
The web is being re-indexed for a new primary user. That user doesn't have eyes. It has an API key.
Sources:
- TechCrunch: Accel-backed Keenable is indexing the web for AI agents
- Keenable: Independent Web Search API for AI
- Andrey Styskin (co-founder): Launch announcement