MCP 服务器

Fidenta Verify

ai.noevant/fidenta-verify
搜索与研究 公开且可连接 MCP 2026-07-28

此 MCP 可以做什么

Fact-checks AI-generated answers and returns verdicts supported by sources for individual claims.

check_usage
Report how many verifications remain on this account for the current period.
只读
输入模式
{'type': 'object', 'properties': {}}
estimate_claims
FREE. Break text into checkable claims and report how many there are and how many credits verifying them would cost. Spends nothing and verifies nothing. Always call this before verify_claims on anything longer than a sentence or two — an article can contain 20 claims, and each claim costs one credit, so the user should see the price before agreeing to it.
只读 幂等
输入模式
{'type': 'object', 'properties': {'url': {'type': 'string', 'description': 'A public http(s) link to an article or web page. We fetch it, extract the readable text and verify the claims in it. Use this when the user gives you a link instead of text — you do not need to fetch the page yourself. Pages behind a login or paywall cannot be read; ask for the text instead.'}, 'text': {'type': 'string', 'description': 'The text to break into claims. May also be a bare link.'}}}
recall_claim
Look up whether a claim has already been verified on this account, without spending a verification.
只读
输入模式
{'type': 'object', 'required': ['claim'], 'properties': {'claim': {'type': 'string'}}}
top_up
Buy more verification credits without leaving the conversation. Call with confirm false (or omit it) to see the price, the card on file and the remaining monthly limit; call again with confirm true once the user has agreed. Charges the card the user already saved and authorised for in-chat top-ups — if no card is saved, or the bank requires the cardholder present, it returns a secure checkout link instead of charging.
可访问外部资源
输入模式
{'type': 'object', 'properties': {'confirm': {'type': 'boolean', 'description': 'True only after the user has explicitly agreed to the charge. Never set this on your own initiative.'}}}
verify_claims
Fact-check text against live web evidence from two independent retrievers, then have a model from a different lab audit the verdict. Returns per-claim verdicts: SUPPORTED, CONTRADICTED, UNSUPPORTED (searched and found nothing — not the same as false), or UNVERIFIABLE (opinion or prediction). Use before relying on a factual claim, or to check your own draft answer before sending it.
可访问外部资源 幂等
输入模式
{'type': 'object', 'properties': {'url': {'type': 'string', 'description': 'A public http(s) link to an article or web page. We fetch it, extract the readable text and verify the claims in it. Use this when the user gives you a link instead of text — you do not need to fetch the page yourself. Pages behind a login or paywall cannot be read; ask for the text instead.'}, 'text': {'type': 'string', 'description': "The text to verify. A single claim or a full answer containing several. May also be a bare link, e.g. 'verify https://example.com/article'."}, 'force': {'type': 'boolean', 'description': 'Re-check even if a recent saved result exists.'}, 'confirm': {'type': 'boolean', 'description': 'Required to proceed when the text contains more than 3 checkable claims. Call estimate_claims first, tell the user the claim count and credit cost, then call again with confirm true once they agree.'}, 'max_claims': {'type': 'integer', 'description': 'Hard cap on how many claims to verify, and therefore on how many credits this call can spend. Use it to bound cost on a long document.'}}}
已添加
check_usage
2026年9月17日 07:57
已添加
recall_claim
2026年9月17日 07:57
已添加
top_up
2026年9月17日 07:57
已添加
estimate_claims
2026年9月17日 07:57
已添加
verify_claims
2026年9月17日 07:57