TL;DR: md2wechat v2.2.0 ships Brand Profile — a local Markdown file that tells your AI Agent who you are, how you write, and what to avoid. Your articles will finally carry your personal voice, not a generic AI tone.

---

§The Problem

More people are using AI to write WeChat articles. But here's the thing: AI output from different people looks the same. Sentence patterns, rhythm, paragraph density — it all blends together.

Every time you start a new conversation, AI treats you as a brand-new user. Your tone preferences, formatting habits, words you hate — you have to explain them all over again.

v2.1.0 solved formatting and de-AI. But one problem remained: AI-generated content lacks your personal signature.

v2.2.0 fixes that.

---

§What's New: Brand Profile

Brand Profile is a Markdown file stored locally at ~/.config/md2wechat/brand.md.

It's not a config file — it's more like a letter to your AI Agent. You tell it: who you are, how you speak, what you dislike.

Two commands to get started:

# Initialize the template
md2wechat brand init

# View current profile content
md2wechat brand show

After initializing, open the file and write three things in your own words:

1. ·Who you are (background, stance)
2. ·How you speak (tone, rhythm, words to avoid)
3. ·Your article preferences (opening style, module preferences)

Once written, every time you ask your Agent to format an article, run md2wechat brand show first. The Agent reads the profile and follows your style.

---

§Why Markdown, Not YAML

Brand Profile is only read by AI Agents, not parsed by the CLI. So the format should be AI-friendly, not machine-friendly.

YAML can say tone: sharp. Markdown can say:

I write like I'm talking to a friend. Lead with the conclusion, then the evidence. I never end with "Hope this helps." I would never start with "In this era of rapid change..."

Counter-examples, specific scenarios, real tone — only natural language captures these. The more specific you are, the more accurately AI reproduces your style.

---

§Formatting Is Reading Decisions, Not Decoration

md2wechat has always been about solving four reader decision points:

Goal	Reader's Question	What Formatting Does
attention	Is this worth reading?	First screen, opening hook
readability	Is this comfortable on mobile?	Paragraphs, steps, tip cards
memorability	What do I take away?	Key quotes, core judgments, author anchors
conversion	What do I do next?	Follow, share, bookmark paths

Brand Profile is the memory layer of this system — letting the Agent know "who this person is" for every article, instead of starting from scratch each time.

---

§What's Next

md2wechat's goal isn't more templates. It's enabling anyone — without design skills — to consistently produce WeChat articles with personal identity.

v2.2.0 delivers the memory layer. Coming next:

• ·Scenario awareness: Auto-select layout strategy based on article type. Tutorials and opinion pieces need different approaches.
• ·Layout Policy: Per-article layout constraints that override brand defaults. Sometimes you want something different for a specific piece.
• ·Markdown Expression Diagnosis: Diagnose plain Markdown and suggest how to present it better. Not editing your content — advising which modules to use.

---

§How to Upgrade

# Homebrew
brew upgrade md2wechat

# npm
npm update -g @geekjourneyx/md2wechat

# Verify version
md2wechat version

No migration needed. Brand Profile is optional and doesn't affect existing workflows.

---

§Final Thoughts

Your articles should look like you — not like "someone using AI."

Brand Profile is step one.

If you're using AI for WeChat content, give it a try. Run md2wechat brand init, spend ten minutes writing your own style profile. You'll notice the output starts sounding like you.

Formatting isn't decoration. It's decision-making on your behalf.

GitHub: github.com/geekjourneyx/md2wechat-skill