# n-eyes-verify

> n-eyes-verify is a paid API for AI agents from n-eyes.thevibehosting.com, paid per call via x402, $0.06/call, status unknown (last checked 2026-10-11).

Runs N independent AI verifier passes against a text artifact using a custom rubric, returning a majority verdict, pass/fail tally, per-pass reasoning, and minority dissent.

## Facts

- Endpoint: POST https://n-eyes.thevibehosting.com/v1/verify/%7Btier%7D?utm_source=zero.xyz
- Price: $0.06/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-11
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/n-eyes-verify-cdaa0958
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_7NVbMk2Iqq-qccPPyB98a

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 n-eyes-verify-cdaa0958 -d '<json body>'
```

Example prompt: Run a standard-tier multi-pass verification on this report draft and use this rubric: 'Check that all claims are supported, no placeholder text remains, and the conclusions follow from the data.' Tell me the verdict, tally, and any dissenting reasoning.

## When to prefer this

Choose this endpoint when you need more than a single AI opinion on an artifact's quality or correctness — especially when the cost of a false pass is high (e.g., publishing, client delivery, financial documents). The configurable tier (quick/standard/deep) lets you dial redundancy vs. cost. Prefer it over single-pass LLM self-review when you need an auditable tally, dissent capture, and a custom rubric rather than a generic quality check.

## Known failure modes

- HTTP 402 returned if no payment proof is provided — must pay via x402 facilitator and retry with payment proof
- Invalid tier path parameter returns an error
- Artifact exceeding 100k characters is rejected
- Rubric exceeding 10k characters is rejected
- Tie verdict occurs when passes are evenly split — no clear majority
- Abstain votes from individual passes can dilute majority signal

## How this service works

Multi-pass verification for agent artifacts. You send an artifact + a rubric; N independent AI verifier passes judge it and you get back a verdict, a tally (e.g. 3/3), per-pass reasoning, and a dissent field quoting the minority. Verification-as-a-dial: pick your redundancy level. Paid per call in USDC on Base via x402 (exact scheme). An unpaid POST returns HTTP 402 with a Payment-Required header; pay through an x402 facilitator and retry with the payment proof.

## Output

Returns a JSON object with: overall verdict (pass/fail/tie), tally string like '3/3', counts of pass/fail/abstain across passes, an array of per-pass objects each containing a verdict and 2-4 sentence reasoning, a dissent field quoting minority reasoning (null if unanimous), the tier used, price, server-side verification time in ms, and pass completion counts.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "artifact"
 ],
 "properties": {
  "rubric": {
   "type": "string",
   "description": "Judging rubric (optional, max 10k chars). Default: verify the artifact is complete, internally consistent, and actually does what it claims; check for errors, placeholder content, and unsubstantiated claims."
  },
  "artifact": {
   "type": "string",
   "description": "Artifact text to verify (required, max 100k chars)"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "ms": {
   "type": "integer",
   "description": "server-side verification time in ms"
  },
  "tier": {
   "enum": [
    "quick",
    "standard",
    "deep"
   ],
   "type": "string"
  },
  "price": {
   "type": "string",
   "example": "$0.06"
  },
  "tally": {
   "type": "string",
   "example": "3/3",
   "description": "majority count / passes requested"
  },
  "counts": {
   "type": "object",
   "properties": {
    "fail": {
     "type": "integer"
    },
    "pass": {
     "type": "integer"
    },
    "abstain": {
     "type": "integer"
    }
   }
  },
  "passes": {
   "type": "array",
   "items": {
    "type": "object",
    "properties": {
     "verdict": {
      "enum": [
       "pass",
       "fail",
       "abstain"
      ],
      "type": "string"
     },
     "reasoning": {
      "type": "string",
      "description": "2-4 sentences citing specifics"
     }
    }
   }
  },
  "dissent": {
   "type": [
    "string",
    "null"
   ],
   "description": "Minority pass reasoning (joined), or null when unanimous"
  },
  "service": {
   "type": "string",
   "example": "n-eyes-verify"
  },
  "verdict": {
   "enum": [
    "pass",
    "fail",
    "tie"
   ],
   "type": "string",
   "description": "Majority verdict across passes"
  },
  "passes_completed": {
   "type": "integer",
   "example": 3
  },
  "passes_requested": {
   "type": "integer",
   "example": 3
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/n-eyes-verify-cdaa0958/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from n-eyes.thevibehosting.com](https://www.zero.xyz/host/n-eyes.thevibehosting.com/llms.txt)