◈

How to automate Xiaohongshu content for marketing teams

Marketing teams chain six local skills into one Xiaohongshu content pipeline: notes, trend scan, brand tone, copy, diagrams, dispatch plan. Data stays local.

Skills in this bundle

1/6

Note Scribe

Create, search, AI-summarize, and export local notes with keyword search across title, content, and tags.

Skill10 Nex
2/6

Content Dispatcher

Deterministic content dispatch plan (by recipient/channel). (Action Co-pilot: generates next-best-actions (actions.json), follow-up tasks and calendar reminders on top of the analysis — what to act on today.)

Skill3 Nex
3/6

Bili Probe

Bilibili content: trending analysis, video info extraction Use case: analysis work in content domains. Costs about 3 Nex per use. Not for unrelated business scenarios.

Skill3 Nex
4/6

Brand Saga

Generate compelling brand stories and narratives for B2C e-commerce brand websites, including brand origin, mission, values, and product storytelling.

Skill3 Nex
5/6

Writing Studio

Content writing production copilot (task-mode x top_n; LLM-orchestrated). (Action Co-pilot: generates next-best-actions (actions.json), follow-up tasks and calendar reminders on top of the analysis — what to act on today.)

Skill5 Nex
6/6

Diagram Smith

Turns 'source -> target: label' lines into a diagram: Mermaid, PlantUML (incl. sequence) or an editable Excalidraw file, with start/end/loop checks.

Skill3 Nex
Quick answer

Run six marketing skills in order inside Nexus Runner and your Xiaohongshu content moves from raw material to a dispatch plan: Note Scribe captures source notes, Bili Probe supplies trend references, Brand Saga fixes the brand voice, Writing Studio drafts the copy, Diagram Smith adds diagrams, and Content Dispatcher groups everything into a per-account, per-channel plan. All six run locally on Windows, and the chain stops at the plan; publishing stays a human step.

  • Six marketing skills form one chain: Note Scribe, Bili Probe, Brand Saga, Writing Studio, Diagram Smith, Content Dispatcher.
  • Everything runs in Nexus Runner 1.0.10 on Windows 10/11 x64; your note library and drafts never leave the machine.
  • One pass through all six costs 27 Nex; the free tier gives 200 Nex per month, reset at the start of each month, enough for 7 passes.
  • No screen-level automation and no third-party account access: the chain ends at the dispatch plan.
  • Each step runs on its own, and Writing Studio produces several structured variants in one call via mode and top_n.

What actually gets automated in a Xiaohongshu workflow

Automation here means fixing the repeatable stages of Xiaohongshu content work, not clicking buttons on the platform for you. Skills take over note capture, trend reference, brand voice, body copy and diagram drafts, and dispatch scheduling. Account login, review, and the act of publishing stay with your team. Every skill in the library is either a skill or an agent; these six are all skills, each with a defined input form and structured output, callable one at a time or chained in the order below. All computation happens inside the local Nexus Runner, so source notes, drafts, and diagrams stay on your own disk.

The six-skill content chain

1Step 1
Step 1: capture material with Note Scribe
Write briefing sessions, user interviews, and comment excerpts into a local note library, tag them, then find them later by keyword across title, content, and tags. It also summarizes the whole library into a structured report and exports Markdown. About 10 Nex per run.
2Step 2
Step 2: pull trend references with Bili Probe
Extract Bilibili trending analysis and video information as a lateral reference for topic angles and headline patterns. It reads Bilibili content, not Xiaohongshu platform data, so treat it as one input among several. About 3 Nex per run.
3Step 3
Step 3: fix the voice with Brand Saga
Generate brand origin, mission, values, and product storytelling for a B2C brand, giving every later post one baseline voice so tone does not drift across writers. About 3 Nex per run.
4Step 4
Step 4: draft copy in bulk with Writing Studio
Combine mode (write, outline, summarize, analyze, rewrite) with top_n to produce several structured drafts, outlines, or summaries in one call. It infers the content type from the input and does not invent missing data. About 5 Nex per run.
5Step 5
Step 5: add visuals with Diagram Smith
Generate Mermaid charts, mind maps, or knowledge hierarchy diagrams as ready-to-edit visual drafts for tutorial and explainer posts, instead of drawing flowcharts by hand. About 3 Nex per run.
6Step 6
Step 6: schedule with Content Dispatcher
Group content deterministically by recipient and channel into a dispatch plan showing which account posts which variant where. This step uses fixed rules with no model call, and it builds the plan only, it does not send anything. About 3 Nex per run.

Three ways to run the same workflow

MCP NexusManual workGeneric cloud AI
Where data and code runSkills run in local Nexus Runner 1.0.10; note library, drafts, and diagrams stay offlineFiles scattered across personal machines and chat threadsSource material and drafts must be uploaded to a third-party server
Path of one postMaterial, trends, voice, copy, diagrams, and dispatch plan fixed as one chainEach stage means another person and another tool, with manual handoffsAssembled ad hoc in a chat session; context is lost in the next one
How you payNex per skill run; 27 Nex for a full six-skill pass, free tier 200 Nex per month reset monthlyCounted in staff hoursPer-seat subscription, same price whether used or not
Bulk output and reuseWriting Studio returns several variants per call; the note library is searchable and reusable by keywordEvery post starts from a blank documentBackground material must be pasted in again each time
Publishing stepChain ends at the dispatch plan: no screen-level automation, no platform account accessEntirely manualAlso has no platform account access

Real use cases

Turn a planning meeting into next week's schedule the same day
Feed in raw meeting notes plus current Bilibili trend data. Note Scribe tags and files the points and summarizes them into a structured report, then Content Dispatcher groups the result into a per-account, per-channel plan.
3 skills in one pass, 16 Nex total
Align brand voice before a product launch
Feed in brand material and product selling points. Brand Saga drafts origin, mission, values, and product storytelling, then Writing Studio expands that baseline into several post bodies and summaries in write mode.
2 skills in one pass, 8 Nex total
Split one long article into several short posts
Feed in an existing long article. Writing Studio produces short variants in rewrite mode, Diagram Smith adds a Mermaid flowchart or mind map, and Content Dispatcher assigns which account posts which variant.
3 skills in one pass, 11 Nex total

FAQ

Can this publish to Xiaohongshu automatically

No. The six skills stop at a dispatch plan. There is no screen-level automation and no access to Xiaohongshu or any other third-party account, so login, review, and publishing stay manual.

Do my drafts and source notes get uploaded to the cloud

No. The skills run inside local Nexus Runner 1.0.10 on Windows 10/11 x64, and the note library, copy drafts, and diagram files stay on your machine.

How much Nex does one full pass cost

27 Nex for all six: Note Scribe 10 Nex, Writing Studio 5 Nex, and Bili Probe, Brand Saga, Diagram Smith, and Content Dispatcher 3 Nex each. The free tier includes 200 Nex per month, reset at the start of each month, enough for 7 passes; Pro is CNY 49 per month with 3000 Nex, Enterprise CNY 299 per month with 50000 Nex.

We do not use Bilibili, do we still need Bili Probe

It is optional. It reads public Bilibili trending analysis and video information as a lateral reference for topics and headlines, not as Xiaohongshu platform data. Skip that step and the other five still chain normally.

Does this require writing code or configuring an API

No. Pick the skill in Runner, fill in the form fields such as Writing Studio's mode and top_n, and run it. Structured output is written straight to local files.

Related use cases