Paste a web page, a post or an email, pick your industry, and see which phrases trip a banned-claim rule and whether the required disclaimer is there. Every flag names the rule and the rule pack behind it. What comes back is a list for review, never a sign-off.
Banned claims and the required disclaimer, from the packs you pick. Every banned-claim rule in each pack, and in the packs it builds on, runs over the text as a fixed pattern (US Regulation D loads 21 rules from three packs). The first pack you pick supplies one required disclaimer, read for its load-bearing phrases. Citation count and banned source domains come back as advisory rows.
Hard is a claim with no legitimate reading, such as a guaranteed return. Review is a phrase with real professional uses, such as "risk-free" beside a Treasury benchmark. Warning depends on context. The packs are researched from the regulators' own text and tested by Verand, not reviewed by a licensed attorney, and drafts are marked as drafts. Your counsel confirms applicability. Not legal advice.
A paraphrase: the attorney rule catches "best personal injury lawyer" and misses "nobody in town wins more cases". Images, video and audio. Whether the piece counts as an advertisement at all, state variations, filing and recordkeeping duties. The extra checks the product adds for the HIPAA, SEC adviser and CPA packs do not run here, and only a pack's default disclaimer is checked.
One request, the rule packs you picked, one pass over the text. The card is the tool in motion on an example sentence we wrote to show what flags look like, and each step lights up while the card is doing it.
The text goes to our server with the packs you picked, and nothing else. A pack brings the packs it builds on: picking US Regulation D also loads US Real Estate Syndication and the Base YMYL Foundation underneath it.
Each rule is a fixed pattern with the regulation it cites. Every occurrence is tested, and a match is set aside only when one of the rule's exception phrases sits within 72 characters of it, so "the risk-free rate" passes and "this fund is risk-free" does not.
The first pack you picked supplies one required disclaimer. The checker reads your text for that block's load-bearing phrases, since you may word it your own way, and flags it with a count when any are missing.
A flag carries its rule id, the pack it came from and its tier. Nothing is rewritten for you and nothing is signed off: what to change is a decision for you and, where it matters, your counsel.
The rules that govern what a regulated business may say about itself are older than the web and stricter than most marketers expect. Here is who writes them, the kinds of claims they catch in each industry, and where a phrase checker helps and where it cannot.
Marketing compliance is the discipline of making sure what a business publishes about itself is allowed by the rules that govern it. For most companies that means one baseline: the Federal Trade Commission Act's ban on unfair or deceptive practices, and the FTC's Endorsement Guides for reviews and testimonials. For a regulated firm there is a second layer on top. An investment adviser answers to the SEC, a broker-dealer to FINRA, a lawyer to a state bar, a doctor to a medical board and to HIPAA, an insurance agent to a state insurance department, and a sponsor raising money for a private deal to the securities laws that govern the offering.
That second layer is why a sentence that is ordinary copy for a coffee shop can be a problem for a financial adviser. "Our clients never lose money" is puffery in one business and a misleading performance claim in another. The consequence is not only a regulator's letter. It is a complaint that names the page, a state bar grievance, or a claim that a private offering used general solicitation when it was not allowed to. The copy is the evidence, and it is public.
No single rulebook covers marketing. Each industry has its own, and each pack in this checker is built from the governing body's own texts. The main ones:
| Who | Rule | What it governs |
|---|---|---|
| Every business | FTC Act Section 5; 16 CFR Part 255 | Deceptive claims and unsubstantiated proof language; endorsements and testimonials, including disclosure of a material connection. |
| Investment advisers | SEC Marketing Rule, 17 CFR 275.206(4)-1 | Every advertisement: seven general prohibitions, plus conditions on testimonials, endorsements, third-party ratings and performance. |
| Broker-dealers | FINRA Rule 2210 | Correspondence, retail and institutional communications; bans promissory claims and performance projections. |
| Private offerings | Regulation D, Rules 506(b) and 506(c) | Whether an offering may be advertised at all, and to whom it may be sold. |
| Lawyers | ABA Model Rules 7.1 to 7.3, as each state adopts them | False or misleading communications, specialist claims, and solicitation. |
| Medical and dental | HIPAA, 45 CFR 164.508(a)(3); FTC health claims | Using patient information in marketing, and substantiation for health and outcome claims. |
| CPAs and tax preparers | AICPA Code of Professional Conduct; Circular 230 | False, misleading or deceptive advertising, and implied influence with the IRS. |
| Insurance | State law built on NAIC model acts; in Canada, CCIR and CISRO guidance | Misrepresenting coverage, benefits or guarantees, and implying government affiliation. |
Two families often appear on marketing-compliance checklists that this checker does not cover, on purpose. Privacy and email laws such as GDPR and CAN-SPAM govern how you collect data and send messages, not what the message claims. And FINRA's rule, although it is in the table, has no pack of its own here yet: the checker has an SEC adviser pack, not a broker-dealer one.
The SEC Marketing Rule treats almost anything an adviser sends to more than one person that offers its advisory services as an advertisement, and any compensated testimonial or endorsement as one too. Extemporaneous live oral communications are outside it. Every advertisement is then held to seven general prohibitions: no untrue statement of a material fact, no statement the adviser cannot substantiate, no misleading implication, no discussion of benefits without fair treatment of the material risks, no cherry-picked specific investment advice, no performance results included or excluded unfairly, and nothing otherwise materially misleading.
FINRA sorts a broker-dealer's written communications by audience: correspondence goes to 25 or fewer retail investors within 30 calendar days, a retail communication to more than 25, and an institutional communication only to institutional investors. Its content standards forbid any "false, exaggerated, unwarranted, promissory or misleading statement or claim", and say a communication may not "predict or project performance" or imply that past performance will recur. Those two lines are why a phrase like "guaranteed returns" sits in the base pack under every financial vertical, not just one.
A real estate syndicator's marketing is governed less by what it says than by whether it may say anything publicly at all. Under Rule 506(b), an offering may take up to 35 non-accredited purchasers who are financially sophisticated, but it may not use general solicitation, which in practice means no public advertising of the deal. Rule 506(c) permits general solicitation, on the condition that every purchaser is accredited and the issuer takes reasonable steps to verify it. The same sentence is therefore fine in one offering and a problem in the other.
That is the logic of the Reg D pack. "Invest now", "limited spots available" and "closing soon" are flagged as active solicitation of a specific offering. "Anyone can invest" and "no minimum net worth" are flagged because Reg D offerings have investor qualification requirements. A return figure such as "earn 14%" is flagged unless a label like "target return", "projected" or "historical return" sits beside it, and "no PPM" is flagged because the offering document is where the risks live. The real estate syndication pack underneath adds the claims that property values always go up, that a syndication interest is liquid, and that tax benefits are guaranteed.
For a medical or dental practice, marketing runs into two separate rules. HIPAA requires a covered entity to obtain the patient's authorization before using or disclosing protected health information for marketing, with narrow exceptions for face-to-face communication and promotional gifts of nominal value; if a third party pays for the marketing, the authorization must say so. That is the rule a patient story, a before-and-after photo or a named testimonial runs into. The second rule is the FTC's: a health claim needs substantiation. The medical pack flags cure and guaranteed-outcome language, "cures arthritis" or "permanent results", as hard, and superlatives such as "painless dentistry" or "top dentist" as review tier.
Model Rule 7.1 bars a lawyer from any false or misleading communication about the lawyer's services, and its comments treat a statement that creates unjustified expectations, or an unsubstantiated comparison with other lawyers, as misleading. The attorney pack encodes those: "we will win" and "guaranteed settlement" are hard, "top trial lawyer" is review tier, a claimed connection with the court is hard, and "no win, no fee" is a warning, because a contingent-fee ad can mislead about who pays the costs; the pack cites Model Rule 1.5(c) and 1.8(e) for it. Model Rule 7.2(c) separately bars a lawyer from stating or implying certification as a specialist unless an approved organization has certified the lawyer and is named in the communication. The pack does not flag "specialist" or "expert" wording today; read those claims yourself against your state's version of the rule.
These are real rules from the packs this checker runs, each with a phrase it flags. They are examples of the pattern, not the whole pack, and every one was tested against the rule it sits beside.
| Pack | Flagged phrase | Rule | Tier |
|---|---|---|---|
| Base YMYL Foundation | “guaranteed returns” | ymyl-guaranteed-returns | Hard |
| Base YMYL Foundation | “you can't lose” | ymyl-cant-lose | Hard |
| Base YMYL Foundation | “risk-free” | ymyl-risk-free | Review |
| US Real Estate Syndication | “property values always go up” | re-guaranteed-appreciation | Hard |
| US Real Estate Syndication | “100% passive income” | re-passive-income-absolute | Warning |
| US Regulation D | “limited spots available” | regd-active-offering-solicitation | Hard |
| US Regulation D | “no PPM” | regd-ppm-not-required | Hard |
| US SEC RIA | “SEC-approved” | ria-bc-003-implied-govt-approval | Hard |
| US SEC RIA | “we're fee-only” | ria-bc-004-fee-only-self-claim | Warning |
| US Attorney / Law Firm | “we will win” | law-bc-002-unjustified-expectations | Hard |
| US Attorney / Law Firm | “no win, no fee” | law-bc-009-pi-fee-cost | Warning |
| US Accounting / CPA / Tax | “former IRS insider” | cpa-bc-005-implied-irs-influence | Hard |
| US Accounting / CPA / Tax | “we're the best CPA in Austin” | cpa-bc-003-self-laudatory | Review |
| US Medical / Dental | “cures arthritis” | hipaa-med-bc-cure | Hard |
| US Medical / Dental | “painless dentistry” | hipaa-med-bc-superiority | Review |
| Insurance (cross-line) | “guaranteed approval” | bi-bc-001-guarantee-beyond-contract | Hard |
| US Health Insurance (draft) | “Medicare-approved plan” | ih-bc-001-medicare-endorsement | Hard |
| US Property & Casualty (draft) | “full coverage” | pc-bc-005-full-coverage-unqualified | Warning |
Each pack carries one comprehensive disclaimer block, and a page is expected to carry it once, not stacked. Firms reword these blocks constantly, so the checker does not demand the exact text. It looks for the phrases that carry the legal weight: for the Reg D pack, "informational", "offer to sell", "accredited investor", "private placement memorandum", "past performance" and "forward-looking". If any is absent, the disclaimer is flagged with a count of how many.
This is also where a careful firm will see its first flag. Below is a real run: the investor-eligibility answer and the Important Disclosures block from the homepage of willowdaleequity.com, a syndication firm run by Verand's founder, pasted as published and checked against the Reg D pack. No banned claim matched. The disclaimer was flagged because one of its six phrases is missing: the block covers offers, the PPM, accreditation, forward-looking statements and past performance, but never calls itself informational. That flag may be right or it may not matter, which is exactly the decision a flag leaves to a person.
A compliance reviewer reads for meaning: who the audience is, whether the piece is an advertisement at all, whether a performance figure is presented fairly, whether a testimonial was paid for. None of that is a pattern. What a deterministic checker does well is the other half: it never tires, never skims the fourth paragraph, and reads every occurrence of every rule the same way every time. It is fast enough to run on every draft instead of the finished one.
The price of that consistency is literalness. A rule written for "best lawyer" does not catch "nobody in town wins more cases", and a rule for "risk-free" fires on "real estate is not risk-free" unless an exception phrase sits nearby. That is why the packs carry exception contexts, why the phrases with real professional uses are review tier rather than hard, and why the first false positive the product fixed was "risk-free" beside the U.S. Treasury benchmark. Treat a clean result as the absence of the listed problems, not the presence of a green light.
Six things that are true of this tool, each one backed by the code that runs it.
The request carries your text and the pack names, and nothing else. There is no account and no database behind the tool, so the copy you paste is checked and then gone.
A banned-claim flag comes back with the phrase that matched, the sentence around it, the rule id, its tier and its pack; a disclaimer flag with how many required phrases are missing. No flag is a bare score.
Fixed patterns, fixed exception phrases, no model in the loop. The same text against the same packs gives the same flags every time, so a result can be re-run and compared.
These are the banned-claims and disclaimer checks Verand runs on every customer draft, against the same bundled packs, called directly. Not a lighter demo version.
The card says who reviewed each pack and marks drafts, and the limits sit beside the result: paraphrases, images and the extra per-industry checks are out of scope.
A run is pattern matching over text you send, so it costs nothing and is never metered. The one limit is 20 checks a minute per visitor.
What it reads, what a flag means, and what it never does with your copy.
Under the SEC Marketing Rule, an advertisement is any direct or indirect communication an investment adviser makes to more than one person that offers its advisory services with regard to securities, plus any testimonial or endorsement the adviser pays for. Extemporaneous live oral communications are excluded. FINRA Rule 2210 uses different words: it sorts a broker-dealer's written communications into correspondence (25 or fewer retail investors within 30 calendar days), retail communications (more than 25) and institutional communications, and holds all of them to its content standards. Deciding which bucket a piece falls in is a judgment this checker does not make, and it has an SEC adviser pack but no FINRA broker-dealer pack yet.
We do not publish frequency counts, so the honest answer is what each pack looks for. Across every financial vertical the base pack flags guarantees, "risk-free", "can't lose" and get-rich-quick language. Syndicators add solicitation ("invest now", "limited spots"), open-to-anyone wording and unlabelled return figures. Advisers add implied SEC approval. Law firms get guaranteed outcomes, "best lawyer" comparisons and "no win, no fee". Medical practices get cure and guaranteed-outcome claims. CPAs get implied IRS influence and "we're the best CPA". The phrase library above lists a real example for each.
No. A clean result means none of the rules in the packs you picked matched and the disclaimer's phrases were found. It cannot tell you the piece is allowed: it does not judge meaning, audience, whether a testimonial was paid for, how performance is presented, or your state's variations. The packs are researched from the regulators' own text and tested by Verand, not reviewed by a licensed attorney. Use the result as a list of things to look at before your compliance reviewer or counsel does, not as their sign-off. Not legal advice.
Fourteen packs, all researched from the regulators' own text and none reviewed by a licensed attorney. Tested by Verand: the Base YMYL Foundation, US SEC RIA / Financial Advisory, US Regulation D, US Real Estate Syndication, US Attorney / Law Firm (a generic state-bar baseline), US Medical / Dental (HIPAA and FTC), US Accounting / CPA / Tax (AICPA and Circular 230), the cross-line Insurance parent, and Canada Life Insurance (CCIR/CISRO and CLHIA). Still in draft, with coverage still being built: US Life, US Health, and US Property and Casualty Insurance, US Fair Housing for apartments, and the Truth-in-Advertising Baseline (FTC and Canada's Competition Bureau) for businesses outside regulated industries. The picker marks drafts.
No. The text travels to our server with the pack names, is matched in memory, and the result comes straight back. There is no account, no database write and no copy kept. The only thing the server counts is how many checks your address has run in the last minute, so it can apply the 20-a-minute limit. Even so, do not paste anything confidential, such as patient details or client names; the check never needs them.
Paid marketing compliance software is built around a review workflow: submissions, approvals, an audit trail and archiving, usually behind a demo and a contract, and mostly for financial firms. This is one deterministic check with none of that, free, across advisers, sponsors, lawyers, CPAs, medical practices and insurance agents at once. Inside Verand the same checks run on every draft the product writes, a hard banned-claim flag stops the draft being published, and a review-tier flag is overridden only by an operator who writes down why.
A phrase checker finds problems after someone has written them. Verand writes articles from your own expertise and credentials, so Google and the AI assistants have something of yours to name. It then tracks where you rank and where ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude mention you, and runs these checks on every draft so a hard flag never publishes.
Content built to rank in
Google and get cited by
ChatGPT
Perplexity
Gemini
Claude, with every claim checked before it goes live.