Underlying tool call
vli_coldseal_readings({
"from": "2026-04-26T00:00:00Z",
"to": "2026-04-26T23:59:59Z",
"out_of_spec_only": true
})
→ returns array of sealed readings
with bundle hashes + registry inclusion proofs VLI exposes your sealed inventory, compliance state, attestations, registry, and live operational data through the Model Context Protocol — so Claude Code (or any MCP-compatible agent) can query it like a teammate.
In conversation
Ask in plain English. Every answer comes back with a sealed tool call underneath.
vli_coldseal_readings({
"from": "2026-04-26T00:00:00Z",
"to": "2026-04-26T23:59:59Z",
"out_of_spec_only": true
})
→ returns array of sealed readings
with bundle hashes + registry inclusion proofs vli_attestations({
"subject_hash": "sha256:abc123..."
})
→ ordered chain of signed attestations
(issuer DID + timestamp + signature) vli_inventory_chain({
"node_id": "47A",
"since": "2026-04-21"
})
→ custody chain with sealed handoff events
per item Catalog
Each tool returns signed evidence. Claude doesn't say it happened — it shows you the sealed record.
vli_dashboardvli_healthvli_incidentsvli_compliance_statusvli_attestationsvli_registry_queryvli_seal_eventvli_inventory_statusvli_inventory_itemvli_inventory_chainvli_checkin_arrivalsvli_patient_receiptsvli_coldseal_nodesvli_coldseal_readingsvli_member_activityvli_members_listInstall
Your SAPI token lives in your Vault session — copy it
into VLI_AUTH_TOKEN. Then add to your Claude Code MCP config:
{
"mcpServers": {
"vli": {
"command": "node",
"args": ["./mcp-server-vli/src/server.js"],
"env": {
"VLI_WORKSPACES_API": "https://your-org.vliworkspaces.com",
"VLI_REGISTRY_API": "https://verifylinkinfra.com/registry-api",
"VLI_SAPI_API": "https://verifylinkinfra.com/sapi-api",
"VLI_AUTH_TOKEN": "your-sapi-token",
"VLI_ORG_ID": "your-org-id"
}
}
}
} Source: github.com/VerifyLinkInfra-cloud/mcp-server-vli · Apache 2.0
Built for orgs
If your org runs Pact, WorkSpaces, TapIn, or Origin, this MCP turns your operational data into a conversation. Your compliance team can ask in plain English. Your auditors can query receipts directly. Your incident response team can check chain-of-custody without opening a dashboard.
Every response is backed by signed evidence. Claude doesn't say the cold chain stayed in spec — it shows you the sealed reading and the registry inclusion proof.
Open source
The MCP server is open source. Fork it, extend it, build your own tools that read from VLI infrastructure. Pull request welcome.
Sealing AI agent decisions instead? AI Trust MCP at /ai/ →
New to VLI? Read the protocol →