All projects matching “fact-check”

Verify factual claims and check whether cited sources support them. Fact-check statistics, dates and source attributions in agent research and reports against web evidence, preserving reporting periods, units and population groups. Returns a verdict, explanation and cited sources; /v2/verify links evidence to source IDs. 0.20 USDC via x402 v2 on Base. Model-assessed; no signup or API key.

For model cascades: is the small model's answer usable, or should the larger model be paid? Detects refusal, evasion, off-topic answers, contradiction and incompleteness; decides accept, review or escalate. Does not fact-check without a source: provide the source to check substance.

Fact-checking with persistent memory — VERITY remembers prior claims your agent has checked and skips repeat lookups, saving tokens across sessions. Returns CURRENT/OUTDATED/DISPUTED/UNVERIFIABLE verdicts with confidence score, sources, and what changed. 0.10 USDC.

Fact-check one claim against live web sources (or URLs you supply). Returns a verdict (supported, refuted, mixed, insufficient), a confidence score and supporting or contradicting quotes that are checked verbatim against the fetched pages. Use it before an agent repeats a claim.

Source-bounded fact-check. Verify a claim against provided source text. Returns verdict: supported, contradicted, or unverifiable. Does NOT search the web.

Fact-check numbers in Korean finance text: every checkable number (rates, KTB yields, FX, listed-company revenue and profit, % vs %p, 조/억/만 units) marked MATCH or MISMATCH against the official value with date and source (ECOS, KEXIM, DART). Numbers with no source stay unverified.

Published fact-checked episodes, newest first, with optional filters by category slug, show id or speaker slug. Paged.

Get the full text of one web page. Give a URL; returns the page's own main content as clean markdown (up to 8,000 characters) with navigation and ads removed, unchanged and not summarized (use /summarize for a short version). Typical response 2 to 5 seconds. If the page cannot be read you get an error and are not charged.

Scan text or a public page for prompt injection before your agent reads it: overrides, fake roles, exfiltration, tool abuse, payment asks, invisible Unicode/HTML-hidden instructions, base64. Rule-based evidence, signed and logged.

Verify an SEC filing: given a filer's CIK and a claimed accession number, does that filing actually exist, and does its form type/filing date match what's claimed? Checked against SEC EDGAR's submissions API. Catches a fabricated or misattributed filing citation, including a real-looking accession number paired with an invented form type or date.

Research cards from KARP, SoulDriver's multi-model research engine: GPT, Gemini, Grok and Claude debate a topic over many rounds and publish a conclusion. Each card gives the conclusion, key insight, warnings, dated figures, consensus status, open questions, research date and staleness. See what is covered first with the free POST /v1/coverage; verify time-sensitive points with POST /v1/check. Cat

x402 seller + agent-readiness audit of one origin: live 402 challenge per paid route (up to 5), price vs catalog, HEAD/garbage-header safety, llms.txt, /.well-known/x402, openapi.json, dead doc links. Read-only.