MCP TRUST INDEX
MCP Trust Index: Launch Edition
The first MCP Trust Index turns Verify's public server evidence into a clear market read: which MCP servers are fresh, compatible, governed, claimed, and ready for serious evaluation before agents call tools.
Search • Sprawl Radar • TrustOps • Gateway • Ledger • Mandates • Hosting • Docs Compiler • Demo • Trust Index • GitHub Action • Plans • Shortlists • Rankings • Compare • Badges • ChatGPT compatible • Claude compatible • Healthcare • Digests • RSS • RSS XML • Backfill • MCP API
Published
May 2026
Launch edition for public MCP trust and buyer readiness.
Primary audience
AI admins
Also useful for platform, security, marketplace, and publisher teams.
Decision lens
evaluation vs production
The index separates client compatibility from production trust approval.
Next action
Use rankings, reports, compare pages, and policy export to review candidates.
Headline read
Public MCP adoption is accelerating, but production trust still depends on fresh validation, authenticated proof, scoped tools, write-action safeguards, and continuous drift monitoring. The first Trust Index frames MCP servers as operational assets, not just connector listings.
- MCP discovery is moving faster than enterprise review capacity, so freshness and drift need to be visible by default.
- Client-compatible is not the same as production-approved. A server can initialize cleanly and still need approval gates.
- Authenticated validation is the key publisher proof point because metadata alone cannot prove scoped behavior.
- Write and exec-capable tools should be routed through policy, mandates, approval, and ledger evidence before production use.
What the index measures
Freshness
Whether validation evidence is current enough for buyer decisions.
Authenticated behavior
Whether OAuth and scoped tool behavior have been proven beyond public metadata.
Tool risk
Whether tools are read-only, write-capable, exec-capable, or approval-sensitive.
Client compatibility
Whether a server looks compatible with ChatGPT, Claude, and generic remote MCP clients.
Production trust
Whether freshness, auth, write safeguards, and policy evidence support approval.
Publisher proof
Whether maintainers can claim, revalidate, monitor, and badge their server.
Buyer actions
| Team | How to use the Trust Index |
|---|---|
| AI admins | Start with fresh evidence, export a policy, and keep stale inventory out of production approval. |
| Platform teams | Use /v1/route to pick candidates and /v1/decide to govern a specific attempted tool call. |
| Security teams | Watch write tools, exec tools, OAuth gaps, freshness drift, and unclaimed publishers. |
| MCP publishers | Claim ownership, run revalidation, publish badges, and move toward authenticated verification. |
Promotion kit
Use this copy and the canonical URL when announcing the launch, linking from docs, or sharing with design partners.
We published the first MCP Trust Index: a buyer-oriented view of public MCP server trust, freshness, client compatibility, authenticated behavior, write-action risk, and production readiness. Use it to separate evaluation-grade MCPs from servers that are ready for governed agent use.
Canonical URL
Shareable proof
Agents can discover the Trust Index through llms.txt and the public MCP surface.
Top 10 index pages
Connector-compatible MCP candidates.
Remote MCP candidates for Claude-compatible clients.
Claims, denial, payer, prior auth, and reimbursement workflows.
Servers with OAuth or protected-resource signals.
Candidates with stronger write and execution safety signals.
Recently validated servers with current evidence.