# Fact Check

> Fact Check is a paid API for AI agents from ZeroClick x402 Service Registry at factcheck.withzero.ai, paid per call via MPP, $0.005/call, status unknown (last checked 2026-10-11).

Verify a factual claim using a second AI model via OpenRouter, returning a structured verdict (true/false/uncertain) with confidence score, reasoning, and corrections.

## Facts

- Endpoint: POST https://factcheck.withzero.ai/run?utm_source=zero.xyz
- Price: $0.005/call
- Payment: MPP
- Status: unknown
- Last checked: 2026-10-11
- Activations on Zero: 0
- Provider: ZeroClick x402 Service Registry
- Website: https://withzero.ai
- Tags: ai, fact-check, verify, llm, accuracy
- Canonical page: https://www.zero.xyz/c/zeroclick-x402-service-registry-fact-check-ad3f3366
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_aIcWQJXgSiCKPGTe9bqHh

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 zeroclick-x402-service-registry-fact-check-ad3f3366 -d '<json body>'
```

Example prompt: Can you fact-check this claim for me — 'The Eiffel Tower is 330 meters tall' — and tell me if it's true or false, how confident you are, and what the correct figure is if it's wrong? Use GPT-4.1-mini for this one.

## When to prefer this

Use this endpoint when you need to cross-validate AI-generated content or user-supplied claims against a second independent model without building your own fact-checking pipeline. Especially useful when your primary model might hallucinate and you want a cheap, structured verdict with reasoning. Prefer a stronger model (e.g. GPT-4.1-mini or DeepSeek V3) for nuanced or complex claims; use the default Gemini 2.5 Flash for speed and cost efficiency.

## Known failure modes

- Claim text is empty or missing — returns 400 validation error
- Claim or context exceeds 5000 characters — returns 400 error
- Invalid model name specified — may fallback or return 400
- Upstream OpenRouter/model timeout — returns 5xx error
- Ambiguous claim with low-confidence result — verdict may be 'uncertain' with low confidence score

## How this service works

Verify factual claims using a second AI model via OpenRouter. Submit a claim and optionally provide context — returns a structured verdict (true/false/uncertain) with confidence score, reasoning, and corrections if the claim is false. Useful for cross-referencing your primary model's outputs against a different model. Default model: Gemini 2.5 Flash (fast and cheap). Use a stronger model for complex claims.

## Output

Returns a structured object with the original claim, the verdict (true/false/uncertain), a confidence score from 0–100, a brief reasoning explanation, corrections if the claim is false (null otherwise), the model used, and response latency in milliseconds.

## Example request

```json
{
 "claim": "Water boils at 100 degrees Celsius at sea level"
}
```

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "claim"
 ],
 "properties": {
  "claim": {
   "type": "string",
   "maxLength": 5000,
   "minLength": 1,
   "description": "The factual claim to verify (e.g. \"The Eiffel Tower is 330 meters tall\")"
  },
  "model": {
   "type": "string",
   "description": "Model to use for fact-checking (default: google/gemini-2.5-flash). Available: openai/gpt-4.1-mini, google/gemini-2.5-flash, meta-llama/llama-3.3-70b-instruct, deepseek/deepseek-v3, mistralai/mistral-small"
  },
  "context": {
   "type": "string",
   "maxLength": 5000,
   "description": "Optional context or source material the claim is derived from"
  }
 },
 "additionalProperties": false
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "claim",
  "verdict",
  "confidence",
  "reasoning",
  "corrections",
  "model",
  "latencyMs"
 ],
 "properties": {
  "claim": {
   "type": "string",
   "description": "The original claim that was checked"
  },
  "model": {
   "type": "string",
   "description": "Model used for fact-checking"
  },
  "verdict": {
   "enum": [
    "true",
    "false",
    "uncertain"
   ],
   "type": "string",
   "description": "Fact-check verdict"
  },
  "latencyMs": {
   "type": "number",
   "description": "Response time in milliseconds"
  },
  "reasoning": {
   "type": "string",
   "description": "Brief explanation of the verdict"
  },
  "confidence": {
   "type": "number",
   "description": "Confidence level 0-100"
  },
  "corrections": {
   "type": "string",
   "nullable": true,
   "description": "Correct information if claim is false, null otherwise"
  }
 },
 "additionalProperties": false
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/zeroclick-x402-service-registry-fact-check-ad3f3366/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from factcheck.withzero.ai](https://www.zero.xyz/host/factcheck.withzero.ai/llms.txt)