The credit bureau for AI agents

One API call. One trust score. Know whether an agent is safe to transact with before a single dollar moves. Built for developers shipping agent commerce.

score.js

                const score = await agentscore.check({

                  agent_id: "agent_9xK2m",

                  action: "purchase_api_credits",

                  amount: 249.00

                });

                // => { score: 87, risk: "low",

                //      flags: [], verdict: "allow" }

How it works

Trust at the speed of agent commerce

Composite trust score

Every agent gets a 0-100 trust score based on transaction history, behavioral patterns, and identity verification. Like FICO, but for machines.

Transaction-level context

Not just "is this agent trustworthy?" but "is this agent trustworthy for this specific transaction?" Amount, action type, and risk signals all factor in.

One endpoint, one response

No SDKs to install, no dashboards to configure. POST a request, get a score. Ship trust scoring in your agent integration in under 5 minutes.

IETF-aligned standards

Built on the emerging trust scoring specification for agent payment transactions. Ready for EU AI Act compliance when enforcement begins August 2026.

Payment rails exist. Trust rails don't. Until now.

AI agents are buying compute, consuming APIs, and managing subscriptions autonomously. The financial plumbing is here. The trust layer was missing. AgentScore fills that gap.