Know What's True. Prove It.
Trace every claim to its origin. Expose distortions, fabrication, and bias.
Fix what's weak — cited or uncited.
See VeraCita for your work
A brief cites a precedent for a proposition the judgment never states. VeraCita pulls the decision, locates the controlling passage, and flags the mismatch — before opposing counsel does.
A regulatory filing cites a standard superseded last year. VeraCita verifies each claim against the current authoritative text and flags what no longer holds.
A literature review reports a study as having “demonstrated” an effect the paper framed as an exploratory association. VeraCita reads the source and grades the overstatement.
A due-diligence memo cites a filing that has since been revised. VeraCita traces the citation to the primary document and flags the superseded source.
An article attributes a statistic to a report that says something subtly different. VeraCita traces the figure to its origin and quotes what the source actually states.
The Core Problem
Three Layers of Verification Intelligence
Surface-level tools check if a source exists. We go deeper — analyzing intent, detecting bias, and finding evidence where none was given.
Detect Distortions & Inaccuracies
Claims that overstate, misrepresent, or twist what sources actually say. A paper says "suggested" — the report says "demonstrated." We catch the difference.
A precedent cited for a rule the court never adopted — the holding stretched one step past the judgment.
A filing cites a standard as it stood before the revision — the obligation has changed, the citation hasn't.
An exploratory association reported as a demonstrated effect — the paper never made the claim the review does.
A figure restated from a filing has drifted — close to the original, but no longer what the document says.
A statistic attributed to a report that measured something subtly different — close enough to sound right.
Expose Bias & Fabrication
Selective citation, cherry-picked data, fabricated references, and hidden conflicts of interest. We analyze not just what's cited — but what was deliberately left out.
The brief cites the one case that helps — and omits the line of authority that cuts the other way.
The favorable data is cited; the adverse findings in the same source are not.
The literature review cites the supporting studies — the contradicting replications never appear.
The deck cites the bullish forecast; the revised guidance from the same source is missing.
The quote is real; the paragraph around it says the opposite of what the story implies.
Source Stronger Replacements
When evidence is weak, missing, or non-existent, we search academic and institutional databases to find credible alternatives. Cited or uncited — we find what matters.
A weak secondary citation replaced with the controlling authority itself.
A blog summary of the rule replaced with the regulator's primary text.
A citation to a press release replaced with the peer-reviewed paper behind it.
A news article's number replaced with the figure from the primary filing.
A second-hand claim traced back and re-sourced to the original document.
VeraCita applies all three layers — automatically detecting distortions, exposing bias, and sourcing stronger evidence.
How It Works
From Upload to Intelligence Report — In Four Steps
Upload any document. Our AI extracts claims, verifies every source, and delivers a comprehensive intelligence report.
Drop Anything In
Upload a report, filing, contract, or thesis — PDF, DOCX, or pasted text. Or drop in AI-generated output from ChatGPT, Perplexity, or NotebookLM. The engine auto-detects whether your content has citations and routes to the right verification strategy.
Any Format
Reports, filings, contracts, theses — PDF, DOCX, TXT, MD, a single claim, or a full report
AI-Generated Content, Too
Works with ChatGPT, Perplexity, NotebookLM — and any other AI output
Cited or Uncited
Auto-detects citations and routes to the right verification strategy — we handle both
We Analyze Everything
Every claim is broken down into its checkable parts, then checked against the cited source and independent evidence. Deterministic checks catch what rules can prove; AI cross-checks catch what needs judgment. Every verdict is calibrated to how much the error would matter.
Claim Decomposition
Every claim broken into its checkable elements, not judged as one block
Evidence From Both Sides
Checked against the cited source and independent evidence — not one alone
Guards + AI Cross-Checks
Rule-based guards and AI review work together; verdicts weighted by how serious the issue is
Review & Strengthen
Every claim gets its own verdict card with the quoted evidence behind it. Edit a claim or swap a source, and it re-checks on the spot. When you're done, export the full verification dossier. AI-assisted, human-directed.
Per-Claim Verdict Cards
Every claim's verdict shown with the exact quoted evidence behind it
Edit-and-Recheck
Change a claim or swap a source, and the verdict updates immediately
Exportable Verification Dossier
A complete, shareable record of every claim, source, and verdict
Get Your Intelligence Report
Annotated documents, source & chain intelligence, compliance audit trails, and more. Up to 13 deliverables per audit — from executive summary to raw data exports.
Annotated Document
PDF, DOCX, or TXT with color-coded verdicts, confidence scores, and inline explanations
Source & Chain Intelligence
Per-source credibility profiles, chain-of-custody reports, and compliance audit trail with digital seal
13+ Deliverables
Executive summary, CSV/JSON exports, session archive (.veracita), and more
The Stakes
The Cost of Getting It Wrong
Misinformation spreads in seconds. AI-generated content looks flawless. In a world where anyone can publish anything, verification isn't optional — it's survival.
Even Legal AI Hallucinates
In Stanford's preregistered evaluation, purpose-built legal AI research tools hallucinated 17–33% of the time — general chatbots, 58–88% on legal queries.12
Today's AI Still Invents Sources
In a 2025 peer-reviewed study, GPT‑4o fabricated nearly one in five citations in literature reviews — and almost half of the real ones contained errors.3
Fake References, Published
A 2026 audit of 2.5 million biomedical papers found AI-fabricated references surged twelve-fold in two years — reaching ~2,900 published papers, almost none corrected.4
Courts Are Acting
Court decisions worldwide documenting AI-hallucinated content in filings — 717 involving lawyers — as of 5 August 2026. Sanctions have reached six figures, alongside disqualification and license suspension.56
The Consulting Wake-Up Call
Deloitte Australia partially refunded a AU$440,000 government report after a researcher spotted AI-invented citations — including a fabricated court quote.7
Finance Is Not Immune
In the first dedicated finance benchmark (2023), the leading AI model answered questions on public filings wrongly — or not at all — 81% of the time.8
Wrong Source, Full Confidence
Columbia's Tow Center found leading AI search engines cited the wrong source in over 60% of news queries — usually without a hint of doubt.9
Malpractice Severity
11 of 13 leading legal-malpractice insurers participated in a claim payout exceeding $100M in two years. At nearly 1 in 4, the average cost merely to defend a claim tops $500K.10
Even Legal AI Hallucinates
In Stanford's preregistered evaluation, purpose-built legal AI research tools hallucinated 17–33% of the time — general chatbots, 58–88% on legal queries.12
Courts Are Acting
Court decisions worldwide documenting AI-hallucinated content in filings — 717 involving lawyers — as of 5 August 2026. Sanctions have reached six figures, alongside disqualification and license suspension.56
Malpractice Severity
11 of 13 leading legal-malpractice insurers participated in a claim payout exceeding $100M in two years. At nearly 1 in 4, the average cost merely to defend a claim tops $500K.10
Fabricated and Misquoted Citations
Citation-level problems documented in court filings — 3,099 fabricated citations, plus misrepresented authority and false quotes — across 1,747 decided cases.5
The Consulting Wake-Up Call
Deloitte Australia partially refunded a AU$440,000 government report after a researcher spotted AI-invented citations — including a fabricated court quote.7
Finance Is Not Immune
In the first dedicated finance benchmark (2023), the leading AI model answered questions on public filings wrongly — or not at all — 81% of the time.8
The AI Act Has Teeth
Non-compliance with the EU AI Act's prohibited-practices rules carries fines up to €35 million — or 7% of worldwide annual turnover, whichever is higher.12
Today's AI Still Invents Sources
In a 2025 peer-reviewed study, GPT‑4o fabricated nearly one in five citations in literature reviews — and almost half of the real ones contained errors.3
Fake References, Published
A 2026 audit of 2.5 million biomedical papers found AI-fabricated references surged twelve-fold in two years — reaching ~2,900 published papers, almost none corrected.4
Retractions Hit a Record
More than 10,000 research papers were retracted in 2023 — an all-time record. The retraction rate has more than trebled in the past decade.11
Finance Is Not Immune
In the first dedicated finance benchmark (2023), the leading AI model answered questions on public filings wrongly — or not at all — 81% of the time.8
The Consulting Wake-Up Call
Deloitte Australia partially refunded a AU$440,000 government report after a researcher spotted AI-invented citations — including a fabricated court quote.7
Today's AI Still Invents Sources
In a 2025 peer-reviewed study, GPT‑4o fabricated nearly one in five citations in literature reviews — and almost half of the real ones contained errors.3
Signed Off Without Evidence
In 2024, 39% of audits inspected by the US audit regulator had deficiencies so serious that the firm signed its opinion without sufficient evidence to support it.13
Wrong Source, Full Confidence
Columbia's Tow Center found leading AI search engines cited the wrong source in over 60% of news queries — usually without a hint of doubt.9
Fake References, Published
A 2026 audit of 2.5 million biomedical papers found AI-fabricated references surged twelve-fold in two years — reaching ~2,900 published papers, almost none corrected.4
Today's AI Still Invents Sources
In a 2025 peer-reviewed study, GPT‑4o fabricated nearly one in five citations in literature reviews — and almost half of the real ones contained errors.3
AI Gets the News Wrong
Journalists from 22 public broadcasters reviewed 2,709 AI answers about the news: 45% contained at least one significant issue — sourcing was the biggest single cause.14
- Magesh, Surani, Dahl, Suzgun, Manning & Ho (2025). "Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools." Journal of Empirical Legal Studies, 22:216–242. onlinelibrary.wiley.com
- Dahl, Magesh, Suzgun & Ho (2024). "Large Legal Fictions." Journal of Legal Analysis, 16(1):64–93. academic.oup.com
- Linardon et al. (2025). "Citation Fabrication and Errors in GPT-4o-Generated Literature Reviews." JMIR Mental Health, 12:e80371 — 35 of 176 citations (19.9%) fabricated; 45% of genuine citations contained errors. mental.jmir.org
- Topaz et al. (2026). Audit of ~2.5M open-access biomedical papers (97.1M references), The Lancet — fabricated-reference rate rose from 1 in 2,828 papers (2023) to 1 in 277 (early 2026); 2,810 affected papers, 98.4% uncorrected. thelancet.com
- Charlotin, D. AI Hallucination Cases Database — 1,846 legal decisions identified, last updated 5 August 2026. damiencharlotin.com
- E.g., Mata v. Avianca, No. 22-cv-1461 (S.D.N.Y. 2023), $5,000; Couvrette v. Wisnovsky, No. 1:21-cv-00157 (D. Or. 2025–26), $110,204 total; Johnson v. Dunn, 792 F. Supp. 3d 1241 (N.D. Ala. 2025), public reprimand, disqualification & bar referral. courtlistener.com
- CFO Dive / Fortune (Oct 2025) — Deloitte Australia refunded the final instalment of a AU$440,000 report for Australia's Department of Employment and Workplace Relations after AI-fabricated references and an invented court quote were identified. cfodive.com
- Islam et al. (2023). FinanceBench (Patronus AI), arXiv:2311.11944 — GPT‑4‑Turbo with retrieval answered incorrectly or refused on 81% of questions about public financial filings. arxiv.org
- Jaźwińska, K. & Chandrasekar, A. (2025). "AI Search Has A Citation Problem." Tow Center for Digital Journalism, Columbia Journalism Review — eight AI search engines, >60% incorrect source attributions. cjr.org
- Ames & Gough (2024). Lawyers' Professional Liability Claims Survey — 13 insurers covering >80% of the AM Law 250. amesgough.com (survey PDF)
- Van Noorden, R. (2023). "More than 10,000 research papers were retracted in 2023 — a new record." Nature 624, 479–481. nature.com
- Regulation (EU) 2024/1689 (EU AI Act), Article 99(3) — administrative fines up to EUR 35 000 000 or 7% of total worldwide annual turnover, whichever is higher. eur-lex.europa.eu
- PCAOB (2025). 2024 annual inspection results — aggregate Part I.A deficiency rate of 39%, down from 46% in 2023. pcaobus.org
- BBC & European Broadcasting Union (2025). "News Integrity in AI Assistants" — 2,709 responses evaluated by journalists from 22 public service media organizations; 45% with at least one significant issue. ebu.ch
Who Benefits
From Cost Center to Competitive Advantage
Stop spending hours on manual checks. A typical 100-citation document completes in under three minutes — cited or uncited.
Traditional Verification
Manual, expensive, error-prone
Endless Burden
Draining resources and morale
VeraCita Advantage
Automated, affordable, accurate
Premium Speed & Accuracy
Competitive advantage delivered
Hours to Days
Time is your enemy
Missing Deadlines
Losing clients and opportunities
Seconds to Minutes
Time is your advantage
Speed Wins Work
First to deliver, first to win
Traditional Costs
Bleeding money on busywork
Manual costs add up per document
Clients question value, margins shrink, losing work to faster competitors
Pay-Per-Use
Massive cost reduction
Massive Savings
Compete on price, take any project size, accelerate growth
No credit card required — Transparent credit pricing, see your balance and usage in real time
When Your Reputation Depends on Accuracy
Legal Professionals
Verify every citation in briefs and filings. Detect misquoted precedents, fabricated references, and selective citation of case law.
Academic Researchers
Verify citations in papers and theses. Check if AI-generated literature reviews hallucinated sources. Detect when "suggested" became "proved."
Financial Analysts
Verify market claims in analyst reports — cited or uncited. Check competitor data, validate statistics, and source stronger evidence for uncited market assertions.
Compliance Teams
Audit regulatory submissions for citation accuracy. Detect selective reporting, verify policy references, and generate compliance audit trails with digital seals.
Consulting Firms
Verify cited sources in deliverables and pitch decks. Fact-check market claims, even when uncited. Source peer-reviewed evidence to strengthen recommendations.
Journalists & Publishers
Fact-check articles and reports — whether sources are cited or not. Detect editorial bias, verify quotes, and expose conflicts of interest behind sources.
Content Agencies
Verify AI-generated content before publishing. Check ChatGPT, Perplexity, and NotebookLM outputs for hallucinated sources and fabricated claims.
Government Agencies
Audit policy papers and legislative research. Verify cited evidence, discover missing data, and trace the chain-of-custody of critical statistics.
Pharmaceutical
Verify citations in regulatory submissions and clinical trial reports. Detect selective reporting of trial results and trace evidence to primary research.
Manual Cite-Checking
Hours per brief, risk at filing
One bad citation in court
Sanctions, disqualification, and reputational damage follow the record
VeraCita on Your Briefs
Checked before it's filed
Defensible briefs
Every citation checked against its source before opposing counsel sees it
Manual Compliance Review
Slow, incomplete, hard to evidence
Findings you didn't catch first
A submission challenged on its evidence becomes a finding against you
VeraCita on Your Submissions
Checked against current primary texts
Evidence you can hand over
A complete record of every claim, source, and verdict
Manual Reference Checking
Late nights before submission
A retraction lasts forever
Fabricated or misrepresented references follow the scientific record
VeraCita on Your Manuscript
Checked before reviewers see it
Defensible scholarship
Every reference checked against its source before submission
Manual QA on Deliverables
Analyst hours on busywork
The client finds it first
One drifted figure undermines the whole deliverable
VeraCita on Your Deliverables
Checked against primary documents
Deliverables that hold
Every claim and source checked before it reaches the client
The Manual Fact-Desk
Deadline versus diligence
Corrections erode trust
Every published error is public and permanent
VeraCita on the Desk
Verdicts before deadline
Checked before it goes out
Claims traced to their origin before the story ships
Features
Built for Verification Intelligence
Every feature engineered to give you complete control over verification quality and cost.
Verification That Thinks
Not just fact-checking. We detect intent, framing, selective citation, and hidden interests behind every source. The engine reads between the lines.
Cited or Uncited — We Verify Both
Upload a cited report or paste a bare claim. The engine auto-detects whether your content has citations and finds the right evidence either way. One platform, any input.
Chain-of-Custody Analysis
Traces every citation back to its primary research origin. Detects citation laundering, broken chains, and context stripping across intermediary sources.
Source Intelligence & Repair
Every source gets a credibility profile — authority score, conflict of interest check, bias indicators, and domain reputation. When evidence is weak, we find more authoritative and credible replacements.
Up to 13 Deliverables Per Audit
Executive summaries, annotated documents, source intelligence reports, chain-of-custody analysis, compliance audit trails, CSV/JSON exports, and more.
Swiss & EU Privacy Standards
Swiss company: we store your data exclusively in Switzerland (AWS Zurich), with AI processing in Switzerland and the EU via AWS Bedrock. In the default private mode your document never leaves your browser — only extracted claim text is sent, browser-direct, to the AI provider. In Server mode uploaded files are deleted once processing completes. GDPR compliant. No training on your data.
Our Own Numbers
Measured, not promised.
Before you trust a verification engine, eight questions matter. We answer each with our own measured runs — real dates, nothing rounded up — and publish the full methodology.
Across every measured run reported above, this engine's errors have gone toward caution, not approval: no trap approved, no true claim condemned. Every figure above carries its run date; the full methodology — including any miss we ever make — publishes with the validation report.
See It In Action
One platform. Three ways to verify. Drop anything in.
Report Citation Checking
"A 200-page due diligence report with 87 citations. Some strong, some outdated, some fabricated."
"A 200-page brief citing 87 authorities. Some controlling, some stretched, one that does not exist."
"A regulatory submission citing 87 sources across studies and standards. Some current, some superseded, some selective."
"A literature review citing 87 papers. Some accurate, some overstated, three the databases have never heard of."
"A 200-page due diligence report with 87 citations. Some strong, some outdated, some fabricated."
"A long-form investigation citing 87 sources. Some solid, some second-hand, some saying something else entirely."
- Upload PDF — engine extracts 87 citations and 142 claims
- Detects 12 distorted claims, 3 fabricated references
- Traces 2 citation laundering chains
- Sources 8 stronger peer-reviewed replacements
- Delivers annotated document + compliance audit trail
- Upload the brief — engine extracts 87 authorities and 142 claims
- Detects 12 stretched holdings, 3 fabricated case references
- Traces 2 chains of secondary commentary to the judgments beneath
- Sources 8 stronger controlling authorities
- Delivers annotated brief + authority audit trail
- Upload the submission — engine extracts 87 citations and 142 claims
- Detects 12 claims resting on superseded texts, 3 fabricated references
- Traces 2 laundering chains back to primary regulatory sources
- Sources 8 current primary texts as replacements
- Delivers annotated submission + regulator-ready audit trail
- Upload the manuscript — engine extracts 87 references and 142 claims
- Detects 12 overstated findings, 3 fabricated references
- Traces 2 citation laundering chains
- Sources 8 stronger peer-reviewed replacements
- Delivers annotated manuscript + verification dossier
- Upload PDF — engine extracts 87 citations and 142 claims
- Detects 12 distorted claims, 3 fabricated references
- Traces 2 citation laundering chains
- Sources 8 stronger peer-reviewed replacements
- Delivers annotated document + compliance audit trail
- Upload the piece — engine extracts 87 sources and 142 claims
- Detects 12 misattributed claims, 3 fabricated references
- Traces 2 chains of second-hand sourcing to the originals
- Sources 8 primary-source replacements
- Delivers annotated article + source audit trail
NotebookLM Verification
"NotebookLM cherry-picks snippets from your sources. Does the snippet actually support the claim? Is there conflicting evidence in the same document it didn't mention?"
- Chrome extension extracts full source documents locally — your data stays private
- Checks whether each snippet actually supports the claim it's attached to
- Scans the full text for conflicting evidence NotebookLM ignored
- Detects selection bias, missing context, and fabricated references
- One-click escalation to full deep audit in VeraCita
Social Media Claim Verification
"Someone posts on LinkedIn: 'Market grew 35% in Q4 2025.' No source. Is it true? Right-click to check."
- Paste a claim or right-click any text on X, LinkedIn, or the web
- Select the platform — engine adapts search strategy accordingly
- Returns verdict: Supported, Contradicted, or Unverifiable with confidence %
- Shows evidence sources with relevance scores
- Share the verified result — first check is free, no account needed
Get the Chrome Extension
Full source documents are extracted locally in your browser — nothing leaves your machine until you choose to verify. VeraCita then checks whether snippets actually support the claims, and scans for conflicting evidence NotebookLM left out.
- Extracts full source texts locally — your data stays private
- Verifies each snippet backs the claim it's attached to
- Surfaces conflicting evidence from the same sources
- One-click deep verification in VeraCita
FOUNDER CIRCLE
The lowest VeraCita rate we will ever offer
500 spots. One price. Your rate locked forever.
“Not every claim has a source. Not every source tells the truth. Not every truth survives the chain of hands it passed through. Now you can check all three — automatically.”
Andy Christen Founder, VeraCita
500 Founder spots — first come, first served
One-time payment. Lifetime access. No subscription required.
- Invite your entire teamSingle shared credit pool. No per-seat fees. Ever.
- Full API accessIntegrate directly into your document workflow. Enterprise-grade endpoints included.
- Founder credits spend across the MindTheGap ecosystemToday, 3 products — every future product included at no extra cost.
- Founder communityExclusive network of forward-thinking professionals. Launching September 2026.
- 10,000 credits per referral (worth CHF 100) — how it worksEarn 10,000 credits (worth CHF 100) for every Founder you refer. Refer 5 colleagues and your entire CHF 500 membership is paid back. Tracked automatically via your unique referral link and credited within 48 hours of each confirmed purchase.
Try VeraCita from CHF 1 · Founder Circle: 30-day money-back guarantee
Return on Investment
A mid-size law firm verifies 10 briefs/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per brief at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Return on Investment
A mid-size law firm verifies 10 briefs/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per brief at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Return on Investment
A mid-size compliance team verifies 10 submissions/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per submission at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Return on Investment
A mid-size research group verifies 10 manuscripts/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per manuscript at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Return on Investment
A mid-size advisory team verifies 10 deliverables/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per deliverable at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Return on Investment
A mid-size newsroom verifies 10 investigations/month, averaging 30 citations each. Manual verification costs CHF 150-400/hour.
Based on 30 citations per investigation at 5 credits per claim. Actual savings depend on document complexity and verification tier selected.
Example: verifying 30 cited claims — 150 credits, CHF 1.50 at the Founder rate. Deeper source audits cost more; the exact price is always shown before you commit.
Founder Perk: AI Transformation Circle
Every competitor can buy the same AI tools. None can replicate your team's values-driven execution. The AI Transformation Circle is a science-based programme that makes AI adoption stick — by starting with people, not technology. Same team, same philosophy as VeraCita: evidence over hype. Founder Circle members get priority access and a preferred rate.
What Comes After — For Those Who Wait
When the Founder Circle sells out, the Subscription plan unlocks at CHF 99/month or CHF 999/year. Pay-As-You-Go stays available at CHF 0.03/credit. No lifetime rate lock. No guarantee. Compare below.
PRICING
Simple, Transparent, One Credit System
No hidden fees. No per-document charges. You see the exact cost before you commit.
- Top up from CHF 20 — use as much or as little as you need
- Credits shared across all MindTheGap products, current and future
Prepaid top-ups from CHF 20 · credits never expire
Credit allowance announced when the plan opens
- Top-up at your plan rate — no surprises
- Set a monthly spending ceiling — never overspend
- Credits shared across all MindTheGap products, current and future
- Up to 5 seats
- Full API access
- GDPR-compliant — DPA available (v1.0)
save 2 months vs monthly · up to 5 seats
vs CHF 0.03 PAYG · locked forever
- 50,000 credits included — never expire
- After exhaustion: top up at the same CHF 0.01/credit rate
- Set a monthly spending ceiling — never overspend
- Credits shared across all MindTheGap products, current and future
- Unlimited seats — no per-user fees
- Full API access
- 10,000 credits per referral (worth CHF 100) · 5 referrals = paid back
- Priority support
- 30-day money-back guarantee — refund minus credits used
50,000 credits included · then top up at your locked CHF 0.01/credit rate
Strictly limited to 500 global spots
Compare all features in detail
| Feature | Pay-As-You-Go Available Now | Subscription Coming Soon | Founder Circle Available Now |
|---|---|---|---|
| Pricing | |||
| Price | CHF 0.03/credit | CHF 999/year or CHF 99/month (credit rate announced at opening) |
CHF 500 one-time (CHF 0.01/credit) |
| Commitment | None — pay as you go | Annual prepaid | One-time, no renewals |
| Credits | |||
| Credits included | Pay for what you use | Announced at opening | 50,000 includedNever expire |
| Credit rate | 0.03 CHF/cr | Announced at opening | 0.01 CHF/cr67% off PAYG |
| Rate locked? | No | No | Yes, forever |
| Credit expiry | None | Announced when the plan opens | Never expire |
| Credits shared across all products of the ecosystem | |||
| Features | |||
| All verification tiers (CHECK / AUDIT / ENHANCE) | |||
| Full audit trail & reports | |||
| Seats | — | Up to 5 | Unlimited |
| API access | — | ||
| DPA (GDPR) | On request | Available | Available |
| Usage dashboard | |||
| Support | Standard | Priority | Priority |
| Founder Exclusive | |||
| Referral commission | — | — | 10,000 creditsworth CHF 100 per referral · 5 = paid back |
| Rate locked forever | No | No | YesCHF 0.01/cr forever |
| Future products | Upgrade required | Upgrade required | All includedAll MindTheGap products, current and future — no extra cost |
| Money-back guarantee | — | — | 30 daysRefund minus credits used · no questions asked |
| AI Transformation Circle | — | — | Priority + preferred rate |
| Founder community | — | — | ExclusiveLaunching September 2026 |
| Get Started | |||
Everything You Need to Know About VeraCita
Get answers to common questions about our AI-powered citation verification platform
Product & Results
Every claim backed by a citation gets verified against its original source. Our AI gives you a confidence score, source quality rating, and a clear verdict: Supported, Partially Supported, Contradicted, Unsupported, or Unverifiable. No more guessing if your citations are accurate.
Around 3 seconds per citation. A 100-citation document processes in under 3 minutes. What used to take 5-8 hours of manual checking now finishes in the time it takes to grab a coffee.
Every verification includes two core deliverables: a color-coded annotated document (green/yellow/red verdicts at a glance) and a detailed verification report with confidence scores and the reasoning behind each verdict. AUDIT and ENHANCE tiers add specialised outputs — source intelligence, chain analysis, audit trails, structured exports — up to 13 deliverables in total. Edit flagged issues in our editor before exporting — no back-and-forth needed.
Any document where accuracy matters: legal briefs, research papers, AI-generated content, consulting reports, regulatory filings. We handle PDFs, Word docs, or pasted text — whatever format your team uses.
VeraCita is an ecosystem of verification tools — all sharing a single credit balance.
Verify every claim, trace the chain of evidence, and strengthen weak sources. Chain analysis reveals whether the evidence behind your citations is trustworthy. Repair mode finds credible replacement sources for weak citations — transforming a flagged report into a defensible one. Color-coded results, confidence scores, in-app editing, and full audit trail.
Two problems with NotebookLM: you can't export citations, and it uses snippets that may cherry-pick from sources. Our Chrome extension extracts all citations and source references automatically. VeraCita then checks the full source text — not just the snippets NotebookLM selected — to detect selection bias, missing context, and fabricated references. Export a clean bibliography or full verified report.
Accounting automation that keeps humans in charge. We handle reconciliation, classification, and matching — so accountants spend time advising clients, not hunting for matching numbers. Take larger mandates, serve more clients. Bank reconciliation, 3-way AP matching, fraud detection. Templates for Swiss and EU firms, with full audit trails.
Your document is parsed in your browser. In Zero-Knowledge mode nothing reaches our servers; in Server mode the extracted claims and source URLs are verified on our Swiss servers, with AI processing in Switzerland and the EU via AWS Bedrock.
Yes. Your credit balance works across the entire VeraCita platform. Whether you're verifying academic citations today or using our upcoming NotebookLM and Financial Audit tools tomorrow, you draw from the same credit pool. Buy once, use anywhere.
Yes. Our engine detects uncited claims and automatically searches academic and institutional databases (CrossRef, OpenAlex, Semantic Scholar) for supporting or contradicting evidence. You get a verdict per claim — Supported, Contradicted, or Unverifiable — backed by sources.
Citation verification audits existing references — checking if cited sources actually support the claims made. Claim verification finds evidence for statements that have no references at all. VeraCita does both automatically — it detects which claims are cited and which aren't, then applies the right strategy.
Yes. Paste any text — a single claim, a paragraph, a full article, or a URL. The engine handles it. Single claims typically get a verdict in under 10 seconds.
Accuracy & Method
Supported: The source backs your claim completely.
Partially Supported: Close, but with caveats — outdated data, missing context, or slight misrepresentation.
Contradicted: The source actively contradicts the claim.
Unsupported: The source does not contain evidence for the claim.
Unverifiable: The source is inaccessible, paywalled, or insufficient to determine a verdict.
Each verdict comes with clear reasoning so you know exactly what to fix.
VeraCita applies the same structured checks without fatigue and shows the evidence behind every verdict. Internal validation results are published on this page — but no automated system is infallible, and expert judgment remains essential in high-stakes contexts.
Yes. We flag when paraphrasing changes meaning, when data is outdated, or when newer research supersedes your source. Protect your credibility before publishing.
Yes. VeraCita handles both cited and uncited documents. For cited content, it verifies claims against their sources. For uncited content, it extracts claims and searches academic and institutional databases to find supporting or contradicting evidence. Either way, you get a full intelligence report.
VeraCita is an AI-assisted verification tool that enhances human review. Our engine catches errors that humans typically miss — but no automated system is infallible. We recommend using VeraCita as part of your professional verification workflow, not as a standalone replacement for expert judgment.
Pricing & Credits
Credit cost depends on your document complexity and verification depth. Simple claims cost 1 credit, complex claims cost 5 credits. Source extraction costs 2 credits per source. Chain analysis, source intelligence, and repair have their own credit costs — all shown as an itemized breakdown before you commit. Founders pay-as-you-go with no monthly minimums. No hidden fees, no wasted spend.
Massive cost savings versus manual verification — pennies per citation instead of dollars per hour. More importantly: avoid the costly risks of incorrect citations — professional sanctions, lost contracts, damaged reputation. A few cents per citation protects years of credibility.
We show you the cost before you commit. Few citations = minimal cost. No citations at all? VeraCita switches from citation verification to claim verification — it extracts the checkable claims and searches for supporting and contradicting evidence directly. Cost then depends on how many claims need evidence discovered.
Credits you purchase never expire — Founder Circle, Pay-As-You-Go, and top-up credits alike. Your Founder rate of CHF 0.01 is locked forever — it never increases, and you pay only for what you use. Subscription plans open after the 500 Founder spots are taken and will state their own credit-validity terms when they launch. All plans show exact costs upfront before you commit.
You can always purchase additional credits at your plan's rate. Founders pay CHF 0.01/credit (same rate as their included credits — forever). Pay-As-You-Go users top up at CHF 0.03/credit, with a CHF 20 minimum top-up. Subscription top-up rates will be announced when the plan opens. You're never locked out — top up only when you need to.
As a Founder, you're not on a subscription — you own your credits outright. After the Founder Circle closes, you can optionally add a subscription for additional monthly credits while keeping your Founder benefits.
All Founder Circle members earn CHF 100 in credits for every Founder they refer. Refer 5 colleagues and your entire CHF 500 membership is paid back. Commissions are tracked automatically via your personal referral link and credited to your account within 48 hours of each confirmed purchase.
Founder Circle members get a 30-day money-back guarantee, no questions asked: a full refund minus any credits you've used, billed at the Founder rate of CHF 0.01/credit. If VeraCita doesn't meet your expectations within 30 days, contact us. We're confident you'll love it.
Workflow & Features
Drag-and-drop or paste text directly. Your team can start verifying in minutes without IT setup. We preserve all formatting and references.
We flag inaccessible URLs instantly. Upload your PDF and our matching engine automatically links it to the right citation using intelligent fingerprinting. Never let paywalls block your verification.
Fix issues in our editor, then export as PDF, HTML report, or CSV. Download audit reports for compliance records. Everything timestamped and traceable.
Export to PDF, HTML, CSV, and JSON formats. Enterprise API available for custom workflows and direct integration into your document management system.
Security & Privacy
Swiss-owned company: we store your data exclusively in Switzerland (AWS Zurich). Your document is parsed in your browser. In Server mode, verification runs on our Swiss servers with AI processing in Switzerland and the EU via AWS Bedrock; in Zero-Knowledge mode nothing reaches our servers. Your data is not used for model training. Uploaded files are deleted once processing completes; verification results are retained so you can return to your session — delete them at any time.
Your document is parsed in your browser. In Server mode, the extracted claims and source URLs are verified on our Swiss servers with AI processing in Switzerland and the EU via AWS Bedrock; in Zero-Knowledge mode they go direct from your browser and never reach us. Uploaded files are deleted once processing completes; verification results are retained for your session until you delete them. Your data is never used for AI model training.
Yes. Our Data Processing Agreement (DPA, v1.0) covers GDPR and Swiss FADP processor terms. Our privacy-first architecture — local document parsing, Swiss-only data storage, deletion of uploaded files once processing completes, and no AI training on your data — makes compliance straightforward.
Complete audit logs: timestamps, AI versions, confidence scores, processing steps. Everything your compliance team needs for regulatory reviews.
Use Cases & Policy
VeraCita serves professionals across multiple domains:
Citation Verification: Legal teams, academic researchers, consultants, journalists, medical professionals, and content agencies — anyone whose reputation depends on accurate, defensible claims.
NotebookLM Citations (coming soon): Researchers and professionals using AI notebooks who need to verify and export their AI-generated bibliographies.
Financial Audit (coming soon): Accountants, auditors, and finance teams handling reconciliation, classification, and compliance across Swiss and EU firms.
Any professional who can't afford errors in high-stakes documents.
Yes. Claim verification supports academic integrity — it's not writing assistance, it's quality assurance.
You remain the author. VeraCita verifies that your sources actually support what you're claiming, traces the chain of evidence behind your citations, and helps you strengthen weak references before submission.
Check your institution's specific policies if uncertain.
No — it keeps humans in charge.
For document verification: AI handles claim checking, chain analysis, and source matching. You decide what to fix and how.
For financial audit: AI handles reconciliation, classification, and matching. Accountants focus on advising clients and making judgment calls — not hunting for matching numbers.
We do the heavy lifting. You make the decisions.
Absolutely — this is a core use case. AI-generated content routinely invents sources and misstates facts — errors that read as authoritative and slip through unless every citation is checked.
Verify catches invented sources, misquotes, and claims that aren't actually supported.
Trace reveals when AI cites sources that themselves rest on weak or untrustworthy foundations.
Repair replaces fabricated or weak citations with credible alternatives — transforming a flagged report into a defensible one.
NotebookLM Citations (coming soon) is built specifically for AI-generated content — automatically extracting and verifying every source from your AI notebook outputs.
Yes. Upload ChatGPT, Perplexity, Claude, or NotebookLM outputs. The engine verifies whether AI-generated citations point to real sources and whether the claims accurately represent what those sources say. We detect AI hallucinations — fabricated references, misquoted findings, and distorted conclusions.
Install the extension and open NotebookLM. The extension extracts citations, source references, and available source text locally in your browser — nothing leaves your machine until you click "Start Free Extraction." VeraCita then verifies whether each snippet actually supports the claim it's attached to, and scans the source text for conflicting evidence NotebookLM may have ignored. Extraction is free; verification costs 5 credits per claim. You can also paste NotebookLM output directly into the VeraCita app without the extension.
Yes. Paste any claim from LinkedIn, X, news articles, or competitor websites. Select the source platform and the engine adapts its search strategy accordingly, returning a verdict with confidence score and evidence sources. Your first verification is free with no account required.
PDF, DOCX, TXT, MD files. You can also paste text directly, enter a URL, or use the Chrome extension for NotebookLM. The engine accepts cited reports, uncited documents, single claims, and AI-generated content.
Enterprise & Scale
VeraCita is built for enterprise workflows. Enterprise API enables custom integrations with your document management systems. Volume pricing scales with your needs — from individual professionals to organization-wide deployment.
Our architecture handles high-volume verification: batch processing for multiple documents, priority queue for time-sensitive work, and full audit trails for compliance. SSO and advanced team features coming soon.
New accounts start with free trial credits — no card needed. When those run out: top up on Pay-As-You-Go at CHF 0.03/credit (CHF 20 minimum top-up). Claim verification costs 5 credits per claim and source extraction 1 credit per source. Deeper source audits cost 13 credits per traced source plus 8 per run — dead-end sources are never billed for tracing, and unused holds are automatically refunded. A typical document verification runs CHF 1–10. You see the exact cost before any processing begins, and you set a spending ceiling so there are no surprises.
Founder Circle members get a 30-day money-back guarantee (full refund minus credits used, at CHF 0.01/credit) if you want to commit upfront.
The Founder Circle is limited to 500 early supporters at a one-time price of CHF 500. All 500 members receive identical benefits:
CHF 0.01/credit — locked forever, including all top-ups. This is 3x cheaper than Pay-As-You-Go and the lowest rate this platform will ever offer.
50,000 credits included — never expire, never reset.
10,000 credits per referral (worth CHF 100) — paid as credits. Refer 5 colleagues and your entire CHF 500 membership is paid back.
All future products included — credits spend across the whole MindTheGap ecosystem, current and future, at no extra cost.
Unlimited team seats, full API access, priority support. 30-day money-back guarantee: full refund minus credits used at CHF 0.01/credit — no questions asked.
VeraCita is open now. Add a card and start verifying from CHF 1 on Pay-As-You-Go. Founder Circle spots are available at CHF 500 one-time — once all 500 spots are filled, the Founder Circle closes and the Annual plan unlocks.
Enterprise clients with high document volumes should contact us at admin@veracita.ai. The Founder Circle is currently the most cost-effective option for any team size — CHF 0.01/credit locked forever, the lowest rate this platform will ever offer.
Yes. API access is included with the Founder Circle and will be available on the Enterprise subscription tier. Documentation is in active development. Contact us if you have a specific integration requirement.
Yes. Any professional email address receives free trial credits upon signup — enough for your first document verifications — with no credit card required. You can verify real documents from your own work before making any decision.
Founder Circle members receive priority support with direct access to the product team during the launch period. Enterprise subscription clients will have dedicated support SLAs. Contact us to discuss your specific requirements.
Still deciding? Verify a document free →
Join the Founders Who Verify Before They Publish
500 Founder spots — first come, first served. Once sold out, the Subscription plan opens at CHF 99/month or CHF 999/year.
Free trial credits on signup, no card required · Founder Circle: 30-day money-back, minus credits used