Kena's web index cuts agent search output tokens by 50%
Token-level scoring over its own index means agents fetch only the passages that matter, which lowers API costs for teams running frequent web searches. Builders wiring up agents through MCP or API can swap in KeenableAI as a search backend and immediately halve the tokens their pipeline pays for. Compared with generic search APIs that return whole snippets, this targets long-document retrieval where most waste happens.
Kena (KeenableAI) runs its own web index for search, giving agents direct web access via MCP or API. The service returns developer-friendly ranked results rather than generic snippets. It splits documents into overlapping chunks, scores individual tokens for relevance to the query, and extracts the most useful passages from long documents. This approach cuts output tokens by around 50% without adding latency.
These guys run their own web index, which makes a big difference when searching the web.
If you have an agent and want to give it web access, you can connect to @KeenableAI (MCP or API). It will return developer-friendly ranked results.
They also built something pretty clever:
Instead of returning generic snippets, they split documents into overlapping chunks, score individual tokens for relevance to the query, and pull out the most useful passages, even when they're buried deep in a long document.
This cuts output tokens by around 50% without adding any latency!
