Know if it's true.
Research questions. Check claims. See the evidence.
ASK produces evidence-backed answers. CHECK evaluates supplied claims and documents. Cortex reports support, qualifications, contradictions, and meaningful limitations.
Ctrl+Enter to run
The problem
AI answers confidently.
Even when it's wrong.
Every major AI system carries a disclaimer: "I can make mistakes." But the answers sound authoritative. People act on them — in medical decisions, financial choices, legal situations.
The problem isn't that AI is always wrong. It's that you can't tell when it is — until it's too late.
"This medication is safe at that dosage."
⚠ Contradicted by FDA drug interaction database
"The market is projected to reach $1.8T by 2030."
⚠ IDC projects $1.3T by 2032 — different figure, different date
"You have 30 days to file a personal injury claim."
⚠ Statute of limitations varies by state — some as low as 1 year
The solution
The authentication layer
for AI-generated information
Cortex retrieves and verifies source candidates, reconciles retained evidence, and reports support, qualifications, contradictions, and evidence gaps with traceable sources.
Authenticated answers
Ask anything. Get answers cross-checked across authoritative sources — not just one AI opinion.
Ask a question →Claim & document analysis
Submit a claim or document for a bounded evidence audit with an evidence-derived Trust Score, verdicts, and stated limitations.
Check a claim →Research-grade reports
The most comprehensive bounded ASK mode, with deeper retrieval, evidence analysis, contradictions, uncertainty, and decision-relevant synthesis.
Start a Deep Report →How it works
From question to verified answer in minutes
Ask or upload
Type a question or upload a document — PDF, Word, or plain text.
Cross-check
Cortex searches authoritative sources and cross-references claims.
Assess & qualify
Material claims are assessed against retained evidence, with qualifications and gaps made explicit.
Decide with context
You receive evidence-derived diagnostics and traceable sources when evidence is sufficient.
Our standard
Cortex doesn't guess.
It checks.
We tell you what we know, what we found, and what we couldn't verify. Transparency isn't a feature — it's the only way trust is built.
Trust requires transparent evidence, traceable sources, stated limitations, and clear acknowledgement when the available evidence is insufficient.
The retained evidence supports the assessed proposition
Evidence supports the core proposition with material limits or scope conditions
Retained evidence materially opposes the proposition under matching scope
The available evidence does not establish the proposition
For everyone
Access to reliable information
is not a privilege
A person in a developing country researching a rare disease deserves the same quality of verified information as a patient at a world-class hospital. Start with Snapshot, then use Cortex Analysis Credits when you need deeper research or verification.
Snapshot
Free to start
Cortex Analysis Credits
From $25
AI told you something important.
Now find out if it's true.
Ask a question or check a claim against real evidence. Start with a free Snapshot—no credit card required. Go deeper with evidence-backed research and document verification when you need it.