{"uid":"cap_dsAybsiWvS0e3q0IQMoRO","slug":"karp-hallucination-check-by-souldriver-a9f50a5f","name":"KARP Hallucination Check by SoulDriver","description":"Hallucination check from SoulDriver: a live fact-check of one claim by four independent AI models (GPT, Gemini, Grok, Claude), each searching the web separately. Returns a summary first: verdict (confirmed, disputed, unverifiable, or contested when the models split), agreement such as 4/4, mean confidence and a one-line headline. Then each model's verdict, reasoning and cited sources. Disagreement is shown, not hidden. No wallet? Free research grant: POST /v1/grant. 10-45 s.","url":"https://agents.souldriver.com.au/v1/check?utm_source=zero.xyz","method":"POST","headers":{},"bodySchema":{"type":"object","properties":{"claim":{"type":"string","maxLength":500,"description":"One factual claim to check"}}},"responseSchema":{"type":"json","example":{"claim":"Spain won the 2026 FIFA World Cup.","summary":{"status":"unanimous","verdict":"confirmed","headline":"Confirmed: all 4 grounded providers agree.","agreement":"4/4","mean_confidence":0.963},"verdicts":[{"verdict":"confirmed","provider":"grok","citations":[{"url":"https://...","title":"..."}],"reasoning":"...","confidence":0.95,"citation_kind":"cited"}]}},"example":null,"exampleRequest":null,"tags":["x402"],"displayCostAmount":"0.25","displayCostAsset":"USDC","priceDynamic":false,"priceHint":null,"priceStatus":"priced","priceSource":"probe","requiresHandshake":false,"reviewCount":0,"rating":{"score":"0.00","successRate":"0.00","reviews":0,"stars":null,"state":"unrated"},"availabilityStatus":"unknown","priceObserved":null,"sessionDeposit":null,"pricing":{"kind":"static","summary":"$0.25/call","primary":{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.25","per":"call","confidence":"exact"},"accepted":[{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.25","per":"call","confidence":"exact"}]},"paymentMethods":[{"uid":"pm_GqFUenaziQVIrSyRVq1n7","protocol":"x402","methodType":"crypto","chain":"base","mode":"charge","costAmount":"0.25","costPer":"request","priority":0,"asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","unit":"request","depositMicros":null,"planRef":null}],"brandName":null,"brandSlug":null,"brandBaseUrl":null,"brandDocsUrl":null,"whatItDoes":"Fact-checks a single claim against the live web using four independent AI models (GPT, Gemini, Grok, Claude) and returns a multi-model verdict with agreement score, mean confidence, and per-model reasoning with citations.","exampleAgentPrompt":"Can you fact-check this claim for me using four different AI models searching the web independently: 'The Eiffel Tower is taller than the Empire State Building'?","exampleUseCases":[{"title":"Verifying AI-generated research facts","prompt":"I just got this output from an AI assistant: 'Lithium-ion batteries were first commercially introduced by Sony in 1991.' Can you run it through your four-model fact-checker and tell me if all the models agree it's correct, along with their sources?"},{"title":"Checking viral news before sharing","prompt":"Before I repost this, can you fact-check the claim 'Spain won the 2026 FIFA World Cup' — I want to see if GPT, Gemini, Grok, and Claude all agree, and what their confidence levels are?"},{"title":"Debunking contested health claims","prompt":"I keep seeing people say 'Drinking eight glasses of water a day is a scientifically proven requirement' — can you run that through your multi-model hallucination checker and show me what each AI model found and whether they all agree?"}],"resultDescription":"Returns a JSON object with a summary block containing: overall status (unanimous/split), verdict (confirmed/disputed/unverifiable/contested), a one-line headline, agreement ratio (e.g. 4/4 or 3/4), and mean confidence across models. Also includes an array of per-model verdicts, each with the provider name, individual verdict, confidence score, natural-language reasoning, and cited web sources with URLs and titles.","failureModes":["Claim exceeds 500-character limit — validation error returned","One or more underlying AI providers are unavailable, potentially reducing agreement denominator","Claim is too vague or ambiguous for models to search effectively, resulting in 'unverifiable' verdict","Timeout beyond 45s if all four providers are slow to respond","Payment failure via x402 if USDC balance is insufficient — endpoint returns 402 before processing"],"whenToPreferThis":"Use this endpoint when you need multi-source, multi-model consensus on a factual claim rather than a single AI's judgment. It is ideal when hallucination risk is high, when publishing accuracy matters, or when a disputed claim needs transparent disagreement surfaced rather than hidden. Prefer it over single-model fact-checkers when you want cited web sources and a quantified agreement score. Not suited for opinion-based or subjective statements.","instructions":null,"reviewSummary":null,"reviewSummaryHighlights":null,"reviewSummaryConcerns":null,"reviewSummaryGeneratedAt":null,"activationCount":0,"lastUsedAt":null,"lastSuccessfullyRanAt":null,"lastHealthCheckAt":"2026-10-30T06:35:27.312Z","isFirstParty":false,"canonicalSlug":"karp-hallucination-check-by-souldriver-a9f50a5f"}