Gera Ground Truth

Live now — you can call this today

The data layer is live as the Gera Verify MCP server: 7 tools backed by real FSA Food Hygiene (10,266 establishments), CQC care registry (5,411 providers) and Gera’s verified-provider set — every answer source-attributed with an as-of date, and it never invents a rating. Hosted endpoint + npm one-line install are shipping next.

Tools an agent can call:

lookup_food_hygiene — real FSA FHRS rating
lookup_care_rating — CQC registration
verify_provider — Gera verified-provider set
check_business_trust — all signals for a business
get_trust_summary — citation-ready sentence

Call it now — two live ways:

# stdio (any MCP client — Claude Desktop, Cursor, …)
npx -y @gera-services/mcp-gera-verify

# hosted HTTP (Streamable MCP)
POST https://gera-verify-mcp-production.up.railway.app/mcp

Open source in the Gera monorepo at packages/mcp-gera-verify (real FSA/CQC data, Ed25519-signed attestations). Both the npm package and the hosted endpoint are live today; a branded verify-mcp.gera.services domain is next — tell us if you want early access.

The gap in the AI age

LLMs are frozen and confidently wrong about anything that changes. An agent that books, buys, or recommends needs a fact it can legally quote with an "as-of" timestamp — not a hallucination.

How it would work

1

Call

An agent hits the API/MCP with a claim ("is this clinic CQC-registered?").

2

Verify

Gera checks it against live licensed sources and the 1M+ listing crawl.

3

Cite

Returns a signed, timestamped, source-attributed verdict the agent can quote.

Who calls it

AI agents & RAG pipelines (per-answer fact checks) and AI labs licensing bulk feeds. Humans almost never — and that is the point.

Why it wins as AI gets stronger

As models get cheaper and more autonomous they take more real-world actions — each one a fact they cannot afford to get wrong. Demand grows with AI capability, not against it.

Call

Verify

Cite

Would you use Gera Ground Truth?