Krews Lab
Where the next
systems start.
Agent architectures, internal tools and prototypes we run on ourselves before anyone else. The parts that hold up end up in Krews Agent and client work.
ProjectStatusExplorePrototypeBuildRelease
Internal toolEvaluation harnessRegression tests for agent behavior. Every production failure becomes a case that runs before the next release.In development
Agent architectureAgent memory runtimeConsolidation, decay and retrieval for agents that need to remember more than one conversation.Prototyping
Internal toolOperator consoleOne place to review conversations, traces and handoffs, and turn bad answers into test cases.Prototyping
Experimental AIAdaptive routingRouting that learns from evaluation results which model handles which request, and at what cost.Exploring
Possible productLog-based crawl analysisRaw server logs turned into a clear picture of how search engines actually spend time on a site.Exploring
Statuses are literal. Exploring means exploring.
Building something adjacent?
If one of these overlaps with a problem you have, some of the work may already exist. Ask.