Verified health claims for AI agents.

An MCP endpoint that returns evidence-backed health claims with sources, safe phrasing, and regulatory flags — so AI agents can publish with confidence.

Verified Claims

Sources

How it works

1. Verify a claim

Send a health claim in English or Japanese. Pramana searches its evidence database for matches.

2. Get evidence + safe phrasing

Receive the evidence level, source citations, quality flags, and regulatory-safe phrasing.

3. Publish with confidence

Use the verified claim in your content. Every response includes a disclaimer, version, and audit trail.

API at a glance

Request

{
  "tool": "verify_claim",
  "input": {
    "claim": "Ashwagandha reduces
      cortisol levels",
    "language": "en"
  }
}

Response

{
  "matched": true,
  "evidence_level": "moderate",
  "safe_phrasing": "Some studies
    suggest ashwagandha may help
    support healthy cortisol levels.",
  "sources": [
    { "doi": "10.1016/...",
      "study_type": "rct" }
  ],
  "yakujihou_safe": true
}

Access tiers

Tier	Rate limit	Tools	Use case
Internal	Unlimited	All tools + admin	YFL services & research agents
Partner	1,000 / day	verify, search, get_claim, list_categories	Approved integrators with API key
Public	100 / day	verify, search (read-only)	Open access with registration
Free	10 / day	verify only	Anonymous / trial