rednote_content_playbook v1.0.0

name: rednote_content_playbook description: "Plan, write, and audit content for Xiaohongshu (RedNote): search-first note strategy, keyword mining, note blueprints for reviews, comparisons, tutorials and itineraries, seeding and KOC patterns, cadence testing, and a compliance audit for absolute claims, fabricated experience, ad disclosure, and off-platform diversion. Use when someone asks how to post, grow, rank in search, or sell on Xiaohongshu or RedNote, or needs a note drafted or reviewed for that platform."

RedNote Content Playbook

Purpose

Xiaohongshu, listed internationally as RedNote, runs on search and peer experience. A post is called a note. Readers open the app to research a purchase, a routine, a trip, or a local visit, then save the notes they plan to act on. That makes the platform closer to a decision log than a follower feed: a small account with the right answer to a typed question can outrun a large account posting general content.

This skill gives an agent a complete working method: mine the exact phrases readers type, pick a note archetype that answers one phrase, draft the note around a save-worthy asset, run the compliance audit, and read results by saves and search behaviour rather than likes alone.

Platform facts in this skill were checked against published sources on 2026-10-11. The source list lives in the production scratch file noted under Operating Rules. Practitioner tactics are labelled as tactics. They are tested practice, not platform guarantees.

Key facts to work from

• Founded in Shanghai in 2013 by Miranda Qu and Charlwin Mao. It began as a guide for Chinese shoppers travelling overseas and grew into a broad lifestyle community covering beauty, fashion, food, travel, home, parenting, study, and local life.
• Reported scale: 300 million plus monthly active users in recent published guides. Audience skew reported consistently at roughly 70 percent female, mostly 18 to 35, with about half in first and second tier cities. Treat these as ranges, not exact current counts.
• Three home surfaces: Follow, Explore (a personalised grid where the cover and title do the work), and Nearby (location based). Search results filter by notes, videos, users, and products.
• Seeding, in Chinese zhongcao or planting grass, is the native conversion pattern: a peer note plants a want, the reader saves it, researches further, and buys later, often inside the app shop.
• KOC means key opinion consumer: an everyday user whose small audience trusts lived experience. KOL means key opinion leader: a larger, taste or expertise led creator. Brands recruit both through Pugongying, the official collaboration platform. Shutiao is the self-serve note boosting tool. Juguang is the larger advertiser platform.
• The save, called a collect, is the strongest public intent signal on a note. A like is a quick reaction. A save means the reader plans to come back, usually before a decision.

Workflow

1. Check fit before any drafting

The platform rewards visual, consumer facing topics a reader would research before acting: a product in use, a routine, a place visited, a meal, a route, a setup, a before purchase comparison. It punishes vague brand broadcasting. If the topic has no searchable question and no real experience behind it, say so and stop. Do not invent experience to fill the gap.

Decide the account posture: personal creator, KOC style consumer account, or brand account. Brand accounts can publish, but peer style notes from real users carry the trust. A brand plan usually needs creator notes through Pugongying alongside the brand's own notes.

2. Mine search demand

Follow references/keyword-worksheet.md. Work inside app search in the language the audience searches in:

1. Write 5 seed phrases in buyer language: condition plus thing plus problem or occasion.
2. Record autocomplete suggestions verbatim, related searches, and the repeated questions in comments under the top notes.
3. Score each phrase with the worksheet table. Cut any phrase the creator cannot back with direct, recent experience.
4. Assign one primary phrase per note. One note, one query. A note chasing three queries usually ranks for none.

3. Pick the archetype

• Honest review: one product or place, real timeframe, pros and a stated drawback, who should skip it.
• Comparison: two to five options, one stated test method, a table. Strongest search format in most consumer categories.
• Tutorial or routine: numbered steps, timing, cost, the step people get wrong.
• Itinerary or store visit: times, locations, transport detail, prices paid, a skip list. Exact logistics drive saves.
• Problem collection: gather the repeated comment questions in a category and answer them in one note.
• Diary with a problem spine: daily life content framed by the searchable problem it solves. Pure diary with no query rarely travels in search.

Full blueprints for three archetypes are in references/worked-examples.md.

4. Draft the note

• Cover: vertical 3:4 image set or vertical video. One clear subject, large readable cover text carrying the shortened primary phrase. Authentic, well lit, real setting. Heavily staged ad creative underperforms peer looking photos as a rule of practice.
• Title: primary phrase at the front, then the deciding detail: the test, the timeframe, the budget, the constraint. Chinese titles are short, so every character must earn its place. English titles get the same discipline.
• Opening two lines: who you are in relation to the problem, and the primary phrase used once in natural language. Readers and search both decide here.
• Body: method before verdicts, specifics over adjectives, prices and timing where you know them, one honest drawback per recommendation, subheadings that carry secondary phrases.
• Save asset: include at least one of a comparison table, checklist, itinerary block, price breakdown, or numbered steps. If a reader has no reason to return to the note, the note is not finished.
• Close: ask a question that invites a specific answer, such as skin type and city, or budget and days. Specific questions produce comments the algorithm and future readers both use.
• Tags: 3 to 8. One broad category tag, the rest mid and long specific phrases from the worksheet. Tag stuffing adds nothing.
• Language: write in the audience's language. Mixed Chinese and English terms are normal on the platform where the product names or category terms are English. Machine translated copy with no native review reads as foreign marketing and underperforms. Flag it for human review when fluency is not available.

5. Run the compliance audit

Run the bundled checker on every draft:

bash
python3 ~/workspace/skills/rednote-content-playbook/bin/note_audit.py draft.md --keyword "oily skin sunscreen" --sponsored

Fix every FAIL before the note goes anywhere. The audit covers:

• Absolute superlatives. China's Advertising Law, Article 9, bans absolute terms in advertising, including best, number one, highest level, and national level, in Chinese or English equivalents. Replace with a specific claim a reader can check.
• Guarantee and universal efficacy language. State your own result, timeframe, and limits instead.
• Off-platform diversion. Links, phone numbers, email, WeChat or WhatsApp IDs, and QR prompts steering readers off platform for sales or follow up are prohibited. Keep contact and purchase paths inside platform tools.
• Fabricated experience. Invented use, fake before and after results, borrowed photos passed off as your own, and undisclosed paid praise are violation categories. If the creator did not do it, the note cannot say they did.
• Disclosure. A paid, gifted, or shopping link note must be identifiable as advertising. Put a plain disclosure at the top and use the Pugongying collaboration flow for brand work.
• Category limits. Financial product promotion requires licensing and platform certification. Medical efficacy claims carry the same risk pattern. Route these topics to qualified, certified accounts or do not publish.

Then read the official community rules before a new account's first post and whenever a note is flagged: https://www.xiaohongshu.com/crown/community/rules

6. Set a test cadence

Cadence figures here are a starting test plan, not a platform rule. Post 3 to 4 notes per week for 4 weeks, each on a different scored phrase, at least two archetypes in the mix. Keep title and cover style steady inside a week so results are readable. Sporadic posting gives search nothing to build on, and daily posting of thin notes trains readers and ranking against the account. After 12 to 16 notes, double down on the archetype and phrase family with the best saves per view, and retire the weakest.

Timing tactic to test, not obey: lifestyle audiences browse heavily at lunch and in the evening. Test one lunch slot and one evening slot in the account's audience timezone and keep the winner based on your own data.

7. Work the first day, then measure the right things

• Reply to early comments with added detail: prices, links inside the platform, the missing step. A useful reply thread extends the note's search life.
• Pin one comment that adds a fact the note missed. Do not pin self praise.
• Record after 7 days: views, saves, comments, save rate (saves divided by views), and whether traffic came from search or Explore if analytics show it.
• Diagnose with the table in references/worked-examples.md. High saves with low comments means make the sequel. High comments with low saves means the next note needs the asset. Good Explore traffic with no search tail means reframe the topic as a typed question.
• A note that keeps gaining search traffic for weeks is an asset. Update it with a follow up note rather than editing the original beyond recognition.

Output Contract

When asked to plan or draft for RedNote, deliver:

1. Fit verdict in one or two lines, including the account posture chosen.
2. Keyword set: one primary phrase, two or three secondary phrases, and the tag list, each traced to the worksheet or marked as untested.
3. Note blueprint or full draft: cover spec, title, opening lines, body with the save asset written out, closing question, tags.
4. Audit result: the note_audit.py output for the draft, with every FAIL fixed or explicitly flagged to the user.
5. Cadence and measurement plan: posting slots to test, the 7 day metrics to record, and the decision rule for the sequel.
6. Disclosure and compliance notes: sponsored status, category limits, and any claim in the draft that rests on the creator's own experience and must be true before posting.

Operating Rules

1. Accuracy first. Every platform fact stated to a user must come from a checked source. Sources for this version: the scratch file at ~/workspace/skill-production/scratch/rednote-content-playbook-sources.md. Numbers that vary by source are given as ranges and labelled reported.
2. Never invent experience, results, prices, or credentials for a note. Placeholders are marked as placeholders and the creator replaces them with real values.
3. No absolute claims, no guarantee language, no off-platform contact or purchase steering in any draft, in any language.
4. Sponsored work gets disclosed in the draft itself, not buried in tags, and goes through official collaboration channels where the account qualifies.
5. Do not scrape the platform, buy engagement, run fake KOC accounts, or post disguised ads. These are the exact behaviours platform enforcement targets, and the account loss is permanent more often than temporary.
6. Respect language and culture fit. If nobody on the team can read the comments the note will attract, say so plainly. Publishing into a conversation you cannot join is a brand risk, not a growth tactic.
7. Keep drafts and audit files with the user. This skill plans and audits content. It does not post to any account on its own.