Copy that sells — and stands up to scrutiny
CopyForge drafts from the facts you provide, then runs transparent text checks before you publish — so common claim and formatting risks are easier to find while you still control the final review.
Why CopyForge exists
CopyForge AI exists because marketplace copy has a recurring problem: avoidable wording such as an invented spec, an unsupported superlative or an unverifiable “best-selling” claim can create listing and advertising-policy risk.
We built a tool that checks copy against a published set of text patterns derived from Amazon, FTC and Meta guidance, then shows the exact matched phrase and source. It is deliberately transparent — and deliberately narrower than a complete legal or marketplace review.
That’s the product: channel-shaped copy with common policy risks surfaced before publication.
Independent on purpose
CopyForge is independently and self-funded. We are building paid plans around useful workflow capacity — not around selling customer data or hiding essential limits.
CopyForge does not sell your data or use generations to train a CopyForge model. Its business model is built around clearly described Pro and Agency subscriptions, not selling customer data.
It’s a simple business model, and it’s why we can be this honest with you on the page you’re reading right now.
Support you can reach
Support is handled directly by the founder who works on the product every day, not an outsourced queue.
If a listing is on the line, tell us it’s urgent so the context is clear. Every support request is reviewed by the person who builds the product — a real human reads every message.
Who builds it
CopyForge AI is designed and run by one person who works on the product every day — not a faceless team or an outsourced queue.
Nabeel Ali
Founder & Brand Designer
Brand designer — 12+ years, 50+ brands across 8 countries. Founder of CopyForge AI.
Before building CopyForge AI, Nabeel spent 12+ years as a brand identity designer for SaaS, fintech, fitness, fashion, real estate and F&B clients across 8 countries — including design roles at Deific and Axact, and training at KIET Karachi and Algonquin College (Canada). That design work taught him the same obsession CopyForge AI encodes into copy: one wrong word, and trust is gone.
How we build
Grounded in your facts, not the model’s imagination
Every generation starts from the product details you paste in, and the prompt instructs the model to stay within them. The deterministic scan then checks configured risk patterns. You must still verify every factual statement: the scanner is not a complete fact-by-fact proof system.
A rule set you can read
The compliance engine is 18 explicit rules, not a black box score. Each one names the policy area it comes from — Amazon listing policy, FTC substantiation and endorsement guidance, Meta ad standards — and each match tells you which phrase triggered it and why. You can disagree with a flag; you cannot be confused about where it came from.
Your copy is yours
Authenticated data queries use row-level security plus explicit ownership filters. Every plan can download its account-linked privacy export; Pro and Agency add spreadsheet-ready copy exports. Account deletion removes primary account content and associated service records as described in the Privacy Policy.
Facts first, with human verification
CopyForge treats your supplied specifications, materials and warranty terms as the permitted fact set. Models can still make mistakes, so the result stays a draft until you verify each statement against your own evidence.
The 18 rules, in full
Every rule the compliance engine runs against your copy, with the policy it is grounded in and the exact phrasing it flags. No black-box score — a list you can read and disagree with on purpose.
Ruleset 2026.09.2 defines 18 explicit text checks. Open a rule for its scope, guidance, source and effective date. The set was reviewed on 2026-09-23; only active rules whose effective date has arrived are applied by the scanner.
Ranking or popularity claimHigh riskamazon · shopify · seo
Amazon restricts promotional/ranking language in listing titles, and popularity claims elsewhere require current evidence. Remove the claim or keep dated, verifiable substantiation.
Read source: Amazon title policy + FTC substantiationEffective from 2026-09-23
Unverifiable social proofHigh riskamazon · shopify · seo · ads · email
Review counts, ratings, customer totals and testimonials must be genuine and supportable. Only publish figures you can evidence from current records.
Read source: FTC reviews and endorsements guidanceEffective from 2026-09-23
Regulated health or medical claimHigh riskamazon · shopify · seo · ads · email
Disease, treatment and health-outcome claims require appropriate competent and reliable scientific evidence and may trigger additional regulatory requirements. Describe substantiated product facts, not an unsupported outcome.
Read source: FTC Health Products Compliance GuidanceEffective from 2026-09-23
Antimicrobial or public-health claim — review requiredHigh riskamazon · shopify · seo · ads · email
Antimicrobial/public-health wording can trigger EPA registration, labeling or treated-article limits in the US, while other markets have their own rules. Do not treat the match as an automatic prohibition: verify the product registration/exemption, target market, exact approved wording and supporting evidence before publishing.
Read source: US EPA treated-articles / antimicrobial claims guidanceEffective from 2026-09-23
Absolute guaranteeMediumamazon · shopify · seo · ads · email
A guarantee must match a policy you actually honour and disclose its material limits. Remove absolute wording that your written terms do not support.
Read source: FTC truth-in-advertising guidanceEffective from 2026-09-23
Scarcity or urgency claimMediumamazon · shopify · seo · ads · email
Deadlines and stock claims must be real and current when displayed. Do not publish static or repeatedly resetting urgency language.
Read source: FTC truth-in-advertising guidanceEffective from 2026-09-23
Unsupported superlativeMediumamazon · shopify · seo · ads · email
An objective comparative superlative implies evidence against the relevant market. Replace it with a specific, supportable differentiator.
Read source: FTC comparative-claim substantiationEffective from 2026-09-23
Absolute durability or lifetime-performance claimMediumamazon · shopify · seo · ads · email
Absolute performance claims ("unbreakable", "lasts forever") require substantiation for the entire product life, which is rarely defensible. State the tested specification or the warranty period instead.
Read source: FTC claim-substantiation guidanceEffective from 2026-09-23
Reference price or discount claimMediumamazon · shopify · seo · ads · email
A former/reference price must be genuine and appropriately recent. Marketplaces may also prohibit pricing language in listing copy.
Read source: FTC Guides Against Deceptive PricingEffective from 2026-09-23
Evidence-dependent safety or certification claimMediumamazon · shopify · seo · ads · email
Keep the test report, specification or issuing-body record that supports this claim. Certification claims should identify the correct standard and must not imply approval beyond its real scope.
Read source: FTC claim-substantiation guidanceEffective from 2026-09-23
Competitor or trademark referenceMediumamazon
Unnecessary competitor marks in an Amazon listing can create accuracy or IP disputes. Use a precise compatibility statement only where permitted and factually necessary.
Read source: Amazon product-title requirementsEffective from 2026-09-23
Off-platform contact or linkHigh riskamazon
Amazon listing copy must not divert a transaction or buyer communication off marketplace. Remove URLs, email addresses and phone/WhatsApp directions from the listing.
Read source: Amazon seller conduct policyEffective from 2026-09-23
Vague emotional fillerLowamazon
This is a quality warning rather than a legal conclusion. Replace low-information filler with a concrete product fact or qualified benefit.
Read source: Copy quality heuristicEffective from 2026-09-23
Before-and-after or transformation claimHigh riskads
Before/after and transformation framing can imply unrealistic or guaranteed outcomes. State what the product is and does, not what the viewer will become.
Read source: Meta Advertising StandardsEffective from 2026-09-23
Implied personal attributeHigh riskads
Ad copy should not assert or imply a sensitive or personal attribute about the viewer. Describe the product or offer without diagnosing the person reading.
Read source: Meta Advertising StandardsEffective from 2026-09-23
Unsubstantiated proof claimHigh riskads · shopify · seo · email
The UK CAP Code requires objective claims to be backed by evidence the advertiser already holds before the ad runs. "Proven" phrasing without a cited study is the classic ASA ruling. State the measurable fact instead, or hold the substantiation ready.
Read source: UK CAP Code (ASA) rule 3.7 — substantiationEffective from 2026-09-23
Unqualified environmental claimMediumamazon · shopify · seo · ads · email
Generic environmental claims without qualification are restricted across the EU under Directive 2024/825 on empowering consumers for the green transition (national measures apply from 27 September 2026). The separate Green Claims Directive proposal, 2023/0085(COD), is still pending and adds no obligations yet — it is tracked separately. Be specific about what makes the product greener ("made from 80% recycled steel") or drop the claim.
Read source: EU Directive 2024/825 (green transition) amending UCPD/CRD; Green Claims proposal 2023/0085(COD) pendingEffective from 2026-09-27
Unsubstantiated lowest-price claimMediumads · shopify · seo · email
Price-comparison superlatives must be true against the market at the moment they are shown and easy for the buyer to check. Unless you run a documented price-match programme, state the price without positioning it against everyone else.
Read source: UK CAP Code rule 3.17 — pricesEffective from 2026-09-23
These are text-pattern heuristics, not a complete policy or legal review. They do not inspect images, product eligibility, landing pages, account history or every category-specific restriction. Verify the current destination policy before publishing.
Ready to test your product catalog?
Start with 10 free daily generations across Amazon, Shopify, SEO, Ads and Email.
Contact & Support
Have a question about marketplace policy, custom agency tiers, or technical integrations? Send us a message and we’ll reply as promptly as we can. You can also email support@copy-forge-ai.com or use the dedicated contact page.