Skip to content
You are offline. Some features may be unavailable.
AI governance evidence platform

Trust, but verify.

CraftedTrust helps teams inventory AI systems, organize vendor evidence, and publish proof that holds up in review. MCP Trust lets buyers compare score, scan depth, and research before connecting a server. Runtime controls and research and CVE discovery extend the same evidence model when review needs to go deeper.

6,975Indexed servers 12Scoring dimensions 2Attestations live 209Coverage live
Aligned to NIST CoSAI 12-factor MITRE CNA (pending) OWASP EU AI Act August 2026
6,975 Indexed servers
54 Average score
2 Attestations live
Apr 26, 2026 Last updated

One evidence model

Inventory

Track which assistants, agents, vendors, and MCP connections exist before governance turns into guesswork.

Evaluate

Combine vendor diligence, public trust signals, and policy context before a system is approved.

Prove

Publish buyer-readable proof, certification status, and ongoing monitoring where it matters.

AIUC-1 Q2 2026

Protocol-aware agent assurance is getting more technical.

The Apr 15 AIUC-1 research update put MCP, A2A, verified identities, logged tool actions, and third-party monitoring closer to the center of buyer review. See what that means in plain English.

MCP publishers start with a free scan

MCP Trust public results | Rate limits apply

Try a live example: https://preclick.ai/mcp

MCP Trust

Public MCP trust signals are one part of the platform.

Use MCP Trust when a buyer needs registry evidence, certification status, and research context before approving an MCP server. Use AI Governance when the decision is broader than a single public MCP connection.

AI Inventory

Map systems, owners, vendors, data access, and approval state in the same evidence model used everywhere else.

Open AI Inventory

Vendor Diligence

Use a cleaner review path for external AI vendors, including MCP providers and higher-risk integrations.

Open Vendor Diligence
Trust note Paid review pays for review work, not a passing result. MCP Trust remains the public evidence layer for MCP. Runtime and governance layers expand the model, but teams still exercise judgment.