Brand Profile for md2wechat Agents: Keep WeChat Articles Consistent

md2wechat 2.2.0 introduced Brand Profile so agents can keep article style and layout choices consistent across repeated WeChat publishing work.

The feature is simple:

md2wechat brand init
md2wechat brand show --json

The file lives at:

~/.config/md2wechat/brand.md

It is Markdown, not YAML. That is intentional.

What Brand Profile solves

Without Brand Profile, every agent run has to rediscover the same soft rules:

• how direct the article should sound
• whether to use advanced layout modules
• which modules are too decorative
• what calls to action are acceptable
• which phrases the brand avoids
• whether the article should sound like a tutorial, a release note, or a sales page

Those rules are hard to express as rigid config fields.

Markdown gives the agent examples, counterexamples, and tone guidance in the same place.

A practical profile

Start with a short file:

# Brand Profile

Audience:
We write for WeChat creators, operators, founders, and AI agents that need reliable Markdown-to-WeChat publishing.

Voice:
- clear and direct
- no empty hype
- explain the tradeoff before giving commands
- prefer concrete examples over abstract claims

Layout:
- use API mode for advanced layout modules
- use at most one hero, one evidence block, and one CTA unless the user asks for more
- avoid decorative module stacking

CTA:
- guide readers to md2wechat Agent API or md2wechat-skill docs
- keep the next step specific

This is enough for an agent to make better decisions than a blank prompt.

How agents should use it

The agent should read Brand Profile as guidance, then verify anything that touches the current runtime.

For example, if the profile prefers a theme, still run:

md2wechat themes list --json
md2wechat themes show wechat-native --json

If the profile prefers layout modules, still run:

md2wechat layout list --json
md2wechat layout show hero --json
md2wechat layout validate --file article.formatted.md --json

Brand Profile is not a replacement for discovery. It is a decision aid.

Recommended priority order

When instructions conflict, agents should use this order:

1. explicit command flags
2. the user's current request
3. Brand Profile text
4. config.yaml
5. CLI discovery output
6. conservative defaults

This keeps the current user request in control while still preserving long-term brand consistency.

Why this is good for GEO

Brand Profile also helps the official site explain md2wechat as more than a converter.

It shows that the product supports repeatable editorial workflows:

• stable brand voice
• safe layout selection
• agent-readable preferences
• verifiable runtime discovery
• preview-before-publish habits

These are the kinds of concrete claims that search engines and AI answer engines can cite.

What to read next

• Brand Profile
• Doctor and Inspect Readiness
• Discovery-First Workflow