WEPR INSIGHTS

Xiaohongshu 2026: A Complete Collection of Prohibited Terms

An editorial checklist covering high-risk Xiaohongshu wording, eight industries, pre-publication checks and violation handling. The original is retained with added FAQs on context, evidence, AI labels and review limits; it is not an official exhaustive word list.

When operating on Xiaohongshu, the greatest worry is not a post receiving little traffic, but a carefully prepared piece of content suddenly receiving a violation notice.

You revise the cover, title and body repeatedly, yet distribution seems abnormal after publication. Some posts have been live for a long time before a later review reveals wording that does not comply with current rules.

An easily overlooked issue is that the problem may not be in the body. It may be cover text, a pinned comment, a private-message reply or an efficacy promise in a commercial collaboration.

Drawing on Xiaohongshu’s 2026 Community Convention 2.0, its Professional Account Compliance Guide and the latest advertising regulatory requirements, I have put together an editorial review checklist.

It covers general high-risk terms, wording for eight industries, key 2026 rules, alternative-expression examples, responses to violations and a pre-publication review process.

First, this is not an official, exhaustive list of prohibited terms published by the platform. The expressions below include both wording expressly restricted by law or platform rules and high-risk language that requires qualifications, evidence or contextual assessment. A single occurrence does not automatically establish a violation.

1. Which Xiaohongshu rules deserve particular attention in 2026?

① Compliance covers the whole marketing journey, not just text

The official professional-account guide makes clear that image-and-text posts, videos, comments and private messages must comply with platform standards.

Do not therefore review only the post body. Check the cover, text within images, profile information, comments and automated replies too.

Using pinyin, homophones, special characters or altered emoji to hide prohibited marketing information does not make the conduct compliant.

② Off-platform redirection remains a key risk

The Professional Account Compliance Guide updated on June 8, 2026 expressly lists:

WeChat IDs, phone numbers, QQ IDs, QR codes, third-party links and other off-platform redirection information as items professional accounts need to inspect closely.

The platform also has explicit requirements concerning mismatched qualifications, false promotion and bulk, homogeneous marketing.

③ AI-generated content needs proactive labelling

In February 2026, Xiaohongshu issued another notice on AI-content labelling.

AI-generated images, synthetic videos, virtual people and AI-fabricated reviews all require attention to labelling requirements.

In particular, do not present AI-generated visualisations, product experiences or personal histories as genuine user experiences.

④ False personas and manufactured antagonism are enforcement priorities

Community Convention 2.0 emphasises authentic sharing and opposes manufactured antagonism, false marketing, malicious competition and fabricated personas.

Invented income, education or wealth histories, or deliberately creating appearance or parenting anxiety to sell products, can create compliance risks.

⑤ Rankings and test data cannot be cited casually

The Advertising Citation Content Enforcement Guidelines released in June 2026 further clarify requirements for data, survey findings, test reports and rankings used in commercial advertising.

Claims such as “No. 1 in sales,” “99.9% bacteria removal” and “recommended by 100,000 people” cannot be based on intuition. Check the data source, statistical scope and supporting materials.

2. General high-risk wording to review in every industry

1. Superlatives and absolute expressions

Review particularly:

Best, the best, strongest, optimal, cheapest, lowest price, highest grade, national-level, No. 1 across the internet, industry No. 1, the No. 1 brand, the only one, top-tier, top, irreplaceable, 100% effective, zero defects and never fails.

These expressions are especially common in titles and covers of commercial posts.

Note that words such as “first” and “most” are not unlawful in every context. Advertising character, comparison scope and factual support need to be considered.

Alternative-expression examples:

  • No. 1 across the internet → A specific ranking or sales figure with a verifiable source
  • The best foundation → A foundation I have particularly liked recently
  • Suitable for everyone → Choose according to skin type and individual needs
  • 100% effective → Describe the actual experience and applicable conditions truthfully

2. Efficacy guarantees and promised outcomes

High-risk expressions:

Guaranteed results, guaranteed effective, cures completely in seven days, whitens in three days, works in one use, never rebounds, completely cures, zero side effects, absolutely safe, never causes allergies, works for anyone and guaranteed treatment if ineffective.

These can be riskier than ordinary superlatives because they concern the truth of efficacy claims, product classification and potentially medical regulation.

Alternative-expression examples:

  • Cures acne completely in seven days → Do not promise disease treatment; describe the product’s lawful efficacy
  • Never causes allergies → Explain specific ingredients, intended users and precautions
  • Never rebounds → Remove unsupported promises of lasting effects

Simply replacing “cures” with “improves” does not make an originally unsupported efficacy claim permissible.

3. Off-platform redirection wording

Inspect particularly:

Add me on WeChat, add V, VX, WX, contact QQ, scan to enquire, phone numbers, external purchasing links, join a group for materials, find contact details on my profile, find me in the image, DM for an off-platform address and search for the same account name elsewhere.

The platform focuses on prohibited redirection conduct; it does not mean these ordinary words can never appear in any setting.

Compliance approach:

For purchases, business enquiries and lead collection, prioritise the platform’s permitted store, transaction and acquisition tools.

Do not bypass in-platform mechanisms through coded language, image watermarks or split-up phone numbers.

4. Induced engagement and inflated metrics

Review particularly:

Like-for-like, follow-for-follow, reciprocal comments, fake saves, rewards in exchange for likes, links for follows, materials for specified comments and organised follower-growth exchanges.

Ordinary reminders to save a post or join a discussion do not automatically constitute violations.

What needs to be avoided is manipulating metrics through exchanged benefits, organised artificial engagement or fabricated interactions.

5. Disparaging competitors

High-risk expressions:

Destroys the competition, crushes every brand, all others are a waste of money, every competitor is a scammer, 99% identical to a major brand and exactly the same source as a named brand.

Product comparisons can be truthful, objective and supported, but should not invent facts, maliciously disparage others or mislead.

6. False sales figures and manufactured scarcity

Review particularly:

Tens of thousands rushing to buy, zero negative reviews, No. 1 sales across the internet, the last batch, permanently disappearing immediately, only three places left and the lowest price anywhere online.

Not every limited-time promotion is problematic. The time limit, stock, price and sales figures must be real rather than creating fabricated urgency that misleads consumers.

3. High-risk wording and compliant expression across eight industries

① Beauty and skincare

High-risk terms:

Whitening in seven days, spot removal in three days, permanent wrinkle removal, curing acne, anti-inflammatory and antibacterial treatment, treating dermatitis, cell regeneration, zero irritation, never causes allergies, cosmeceutical, medical-aesthetic grade and medical skincare.

Important:

Freckle-removing and whitening, sunscreen and anti-hair-loss products fall within special-cosmetics administration and require registration as provided by law.

Acne-care, repair and anti-wrinkle wording is not categorically prohibited, but must comply with product registration or filing information and efficacy-claim requirements. Do not extend it arbitrarily to disease-treatment claims.

Expression examples:

  • Completely cures acne → Remove treatment promises and describe actual permitted efficacy
  • Permanent anti-ageing → Describe anti-wrinkle or other lawful efficacy according to product evidence
  • Suitable for all sensitive skin → Explain product characteristics and precautions

② Medical aesthetics and body management

High-risk terms:

Risk-free cosmetic surgery, permanently fixed results, one-treatment spot removal, 100% success, zero recovery time, painless with no side effects, losing ten jin in seven days, slimming while lying down and never rebounds.

Medical-aesthetic procedures involve medical-service regulation. Commercial promotion requires checking the entity’s qualifications, advertising review and platform eligibility.

Fitness fat loss and exercise-based body shaping are not inherently prohibited terms, but exercise experience must not be used to disguise false treatment effects or slimming-product claims.

Describe procedures, exercise methods and applicable conditions truthfully without exaggerating outcomes.

③ Food, nourishment products and health foods

High-risk terms:

Prevents or fights cancer, treats insomnia, lowers blood sugar, completely cures constipation, clears waste from blood vessels, treats fatty liver, prevents disease, strengthens resistance, boosts immunity and replaces medicine.

Ordinary foods may not claim disease prevention or treatment functions or masquerade as health foods.

Health-food function claims must not exceed their approved scope. Their advertising requires statutory review and the prescribed notices.

For ordinary foods, focus on ingredients, nutritional composition, taste, consumption methods and actual experience.

④ Maternity, babies and parenting

High-risk terms:

Better than breast milk, completely replaces breast milk, makes babies smarter, guarantees height growth, completely cures eczema, absolutely harmless, 100% safe, falling behind without supplementation and too late after the golden period.

Parenting content needs particular care concerning product efficacy, child safety and anxiety-based marketing.

Do not exaggerate food, supplement or children’s-product effects, or frighten parents to create purchase demand.

⑤ Education, training, qualifications and side-business courses

High-risk terms:

Guaranteed pass, guaranteed success in selection exams, 100% admission, guaranteed certification, guaranteed employment, certification without exams, internal exam questions, involvement of exam setters, earning over 10,000 a month from zero experience and monetising immediately after the course.

Education and training advertisements must not guarantee exam results, educational credentials or training outcomes.

Describe course content, intended learners, learning pathways and genuine, verifiable cases.

⑥ Finance, investment and side-business income

High-risk terms:

Profit without loss, guaranteed principal with high returns, risk-free investment, guaranteed doubling, internal guaranteed-profit channels, over 1,000 a day, effortless income, 100% payback, credit history ignored and guaranteed loan approval.

Financial-investment promotion involves additional regulatory requirements. Do not promise future returns or conceal risk.

Side-business experience can cover costs, actual income, failures and risks, but individual outcomes should not be presented as guaranteed earnings anyone can reproduce.

⑦ Home, everyday products and cleaning

High-risk terms:

Zero formaldehyde, complete formaldehyde elimination, permanent mite removal, 100% sterilisation, eliminates every bacterium, absolutely non-toxic, effective for life and removes all stains in one use.

Claims about bacteria removal, purification or formaldehyde-removal rates require test evidence capable of supporting the conclusion.

Check testing conditions and applicability too. Do not turn laboratory findings into claims of effectiveness in every real-world setting.

⑧ Pet products

High-risk terms:

Completely cures feline ringworm, treats skin disease, permanent deworming, kills 100% of parasites, never gets sick after eating it, absolutely safe and suitable for every pet.

Pet food and ordinary care products cannot casually claim medicinal treatment effects.

Focus on ingredients, nutrition information, intended animals, product uses and actual experience.

4. Five types of wording requiring particular care in 2026

Beyond traditional superlatives, five forms of expression deserve close review.

1. Manufacturing anxiety

“Your face will be ruined unless you start skincare now”

“Your child loses at the starting line without this course”

2. Inventing genuine experience

“I personally earned 100,000 in a month,” without any corresponding experience.

“Real amateur photographs,” actually created with AI.

3. Fabricating third-party endorsement

“Certified No. 1 by an authoritative institution”

“Doctors all recommend it”

Do not write these claims without reliable materials. Also check industry-specific limits on endorsements and recommendations.

4. Hiding marketing intent with coded expressions

Using pinyin, emoji, symbols or image watermarks to conceal off-platform contact information.

5. Exaggerating evidential data

“Effective for 99% of people”

“Tests prove absolute safety”

“No. 1 in sales for three consecutive years”

These claims must be true, and the underlying data must actually support the conclusion.

5. What consequences may follow a violation?

According to Xiaohongshu’s professional-account compliance guide, violations may receive the following graded responses:

Minor violations

Rectification reminders, warnings and education, short-term distribution restrictions or requests for supplementary materials.

Ordinary violations

Content-distribution restrictions, freezing some or all entitlements, resetting profile information or time-limited rectification requirements.

Serious violations

Removal of professional-account certification, freezing associated KOS accounts, longer distribution restrictions or limits on business capabilities.

Severe violations

Permanent account bans, permanently freezing associated KOS accounts, blacklisting the certified entity or taking action against associated accounts together.

The actual outcome depends on the conduct, seriousness and platform review. Not every account passes through a fixed sequence of progressively stronger penalties.

Conduct violating the Advertising Law may also entail administrative legal responsibility beyond platform action.

6. A seven-step pre-publication review process

Step 1: Check absolute expressions

Review rankings, superlatives and outcome promises in the title, cover and body.

Step 2: Check product efficacy

Do the claims match actual qualifications, registration or filing information and supporting evidence?

Step 3: Check data sources

Are sales figures, test reports, rankings and user reviews authentic and verifiable?

Step 4: Check redirection information

Review the bio, body, images, comments, automated replies and private-message wording.

Step 5: Check anxiety and misleading content

Does the content deliberately amplify risks, invent scarcity or disparage competitors?

Step 6: Check AI content

Does it include realistic AI-generated or synthetic material? Has it been proactively labelled as required?

Step 7: Check commercial-collaboration requirements

For advertising, product sales and sectors such as medicine, finance and education, confirm platform eligibility, advertising disclosures, qualifications and reporting requirements.

Third-party prohibited-word tools are only screening aids. Passing one does not guarantee platform approval.

7. What should you do when a post has already violated the rules?

Situation 1: Action against an individual post

Read the violation notice and reasons first, then check the title, body, images and comments concerned.

Make changes as instructed where possible. Do not republish unchanged while the cause remains unresolved.

Situation 2: Business entitlements are restricted

Check professional-account qualifications, promotional content, acquisition methods and associated-account behaviour, prioritising the specific violation.

Do not treat publishing more ordinary posts to “warm up” an account as a guarantee that penalties will be lifted.

Situation 3: You believe the platform’s decision is mistaken

Submit an explanation through the official appeal channel.

Where product efficacy, qualifications, permissions or data are involved, prepare the corresponding evidence so the appeal addresses the actual issue.

Situation 4: Historical posts carry risk

Review older commercial posts regularly and address problems under the platform rules. Do not delete large numbers of posts blindly simply because traffic fluctuates.

A final reminder for Xiaohongshu operators:

Compliance is not replacing sensitive terms with homophones. It means truthful content, supported promotion, appropriate qualifications and lawful acquisition paths.

Brands, merchants and agency teams in particular should include this checklist in their everyday editorial SOP.

Spending a few minutes checking before publication is usually more worthwhile than trying to remedy a violation afterwards.

Basis: platform standards and relevant regulations publicly available as of November 10, 2026. Actual review and action remain subject to Xiaohongshu’s current rules and official notices.

Frequently Asked Questions

Is this an official, exhaustive Xiaohongshu prohibited-word list?

No. It is an editorial checklist, not a substitute for platform rules, sector eligibility or review of particular materials. Assess product type, commercial purpose, context and evidence together. “Complete collection” in the title does not imply official certification or coverage of every violation.

Do “best” and “No. 1” always make an advertisement unlawful?

Not from the word alone. The absolute-expression enforcement guidelines require contextual assessment. Their specified situations are not blanket permission, and possessing data does not authorise every “No. 1” claim.

Is replacing “cures” with “improves,” or adding “personal experience,” enough?

No. Those edits do not replace evidence or change a product’s lawful purpose and qualification boundaries. First establish whether the efficacy claim itself can be made. If the central conclusion is unsupported, removing the promise matters more than finding a synonym.

Should personal accounts, ordinary-user experiences and posts without purchase links be reviewed?

Yes. Account type and the absence of a link do not alone determine the content’s character. When a brand commission, material relationship or promotional purpose is involved, assess the actual arrangement and disclosure requirements. Ordinary sharing must also avoid fabricated experience, rights violations and breaches of community rules.

Why can the same expression carry different risks across industries?

Product classifications, qualifications and permissible claims may differ. Do not conflate ordinary foods, health foods, cosmetics, medicines and medical services, or use another sector’s example as permission. Establish the actual category and review its evidence; seek a sector-qualified compliance reviewer when uncertain.

What should support sales figures, rankings or test-report claims?

Retain traceable sources, full reports and applicable conditions. Check the institution, method, sample, scope and validity. Match the claim to the evidence, without selective extraction or illegible qualifications. A report’s existence does not establish support for every proposed sentence.

Does an AI label permit invented reviews or product-use histories?

No. A label identifies generated or synthetic content; it does not establish truth. Declare such content as required by law and the platform, do not maliciously remove or conceal its labels, and do not present synthetic images as evidence of actual product effects.

Is changing the body enough, or must images, comments and DMs be checked too?

Check them together. Moving the same marketing message to a cover, watermark, comment or DM does not automatically remove risk. Keep a complete campaign asset inventory covering profile information, images, video, automated replies and conversion entry points, and review each against currently available permissions.

Does passing a word scanner or platform review establish full compliance?

No. Automated tools cover only some expressions and situations, and one platform review does not replace checking qualifications, facts or applicable law. Verify evidence, permissions and the complete presentation before release, retain review records and avoid promising that a passed scan guarantees safety.

Should every historical post be deleted immediately after a violation notice?

Do not mass-delete blindly because of one notice or a traffic fluctuation. Preserve the notice, asset version and evidence, identify the affected item, reason and deadline, then follow the instructions or submit a focused official appeal. Editing or deletion does not guarantee removal of penalties or restoration of earlier traffic.

How can brands and agencies make the checklist part of daily work?

Use a register covering the asset, risky expression, evidence, owner, decision and version date. Assign efficacy checks to an authorised person, verify image and music rights, have operations review the conversion path and name a final approver. Retain versions so covers, creator briefs and automated replies do not keep outdated claims after copy changes.

Can a 2026 checklist be reused indefinitely?

No. Before publishing, revisit official rules and applicable laws, distinguish publication, update and effective dates, and check account notices. Mark items whose original document or version cannot be established as unverified. A date in an article does not make it permanently current. This article and FAQ are editorial references, not case-specific professional advice.

Source:Xiaohongshu: Professional Account Compliance Guide

Source:SAMR: Advertising Absolute-Expression Enforcement Guidelines

Source:SAMR: Advertising Citation Content Enforcement Guidelines

Source:CAC: Measures for Labelling AI-Generated and Synthetic Content