About
Verging Labs is an independent performance lab for AI agents. We measure what every tool actually adds, and we publish the results.
Why This Exists
An agent's intelligence is assembled: the model, plus the tools it works through. The model half gets all the attention. The tool half decides how much of that intelligence actually reaches your users, and it goes largely unmeasured. So most agents run below their potential, and the people running them can't see by how much.
This project started when I watched usage-ranked leaderboards in the x402 ecosystem get gamed in real time. The fix is boring and hard: measure continuously, publish the methodology, and never sell placement.
Matt Overing
What Is Live and What Is Next
- Live now: the Agentic Search Index v0.1, judge-calibrated with the agreement statistics published on the methodology page, and the telemetry layer across 7 categories.
- Next: the Agentic Memory Index.
Who Runs It
Built and run by Matthew Overing. The measurement infrastructure and this site are operated as a single independent effort. There are no provider investors and no paid listings.
Questions, corrections, or a tool we should benchmark: contact@agenticresourceradar.com.