# VerityLayer Hallucination Verification Engine

> VerityLayer Hallucination Verification Engine is a paid API for AI agents from api.veritylayer.dev, paid per call via x402, $0.25/call, status unknown (last checked 2026-09-16).

Verifies a claim or AI-generated statement against authoritative sources and returns a verdict, evidence, reasoning, and confidence score to catch hallucinations before they reach users.

## Facts

- Endpoint: POST https://api.veritylayer.dev/verify
- Price: $0.25/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-16
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/veritylayer-hallucination-verification-engine-faadbb39
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_3nB9a35v7ahRxnCGyianD

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability veritylayer-hallucination-verification-engine-faadbb39 -d '<json body>'
```

Example prompt: Before we send this response to the user, run it through VerityLayer to check if the claim 'The Eiffel Tower was built in 1889 on the Champ de Mars in Paris' is actually supported — I need the verdict, evidence sources, and confidence score.

## When to prefer this

Choose VerityLayer when you need a fail-closed, independent fact-checking layer specifically designed to catch LLM hallucinations — especially fabricated citations, fake statistics, or unsupported claims — before they reach end users. It is particularly well-suited for legal, medical, research, or customer-facing AI applications where a single hallucination can cause serious harm. Prefer this over general web search or RAG retrieval when you need a structured verdict with evidence and confidence rather than raw search results.

## Known failure modes

- Claim too vague or ambiguous for authoritative lookup — may return low-confidence result
- No authoritative sources found for niche or very recent claims — may return 'uncertain' verdict
- Overly long or compound claims may reduce accuracy — works best with atomic, single-fact statements
- Network or upstream source availability issues may cause timeouts or errors
- Rate limits or payment failures may result in 4xx/402 responses

## How this service works

VerityLayer verifies whether a factual claim is true — fact-check, verify a claim, is this true. Independent, fail-closed verification for AI agents: returns a verdict (supported/unsupported/uncertain), a calibrated 0-1 confidence, reasoning, and evidence; abstains rather than guess. Every verdict includes an Ed25519-signed receipt — portable cryptographic proof of what was verified, free re-verification at /receipt/verify. Tier: grounded — live web citations (Sonnet, default).

## Output

A JSON object containing: 'verdict' (e.g. 'supported', 'unsupported', 'uncertain'), 'evidence' (array of authoritative source citations supporting or refuting the claim), 'reasoning' (a human-readable explanation of how the verdict was reached), and 'confidence' (a 0–1 numeric score indicating certainty of the verdict).

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "claim": {
   "type": "string"
  },
  "context": {
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "verdict": "supported",
  "evidence": [
   "Encyclopaedia Britannica — Eiffel Tower, Paris, France."
  ],
  "reasoning": "Multiple authoritative sources place the Eiffel Tower on the Champ de Mars in Paris, France.",
  "confidence": 0.97
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/veritylayer-hallucination-verification-engine-faadbb39/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.veritylayer.dev](https://www.zero.xyz/host/api.veritylayer.dev/llms.txt)