Four signs your AI post sounds fake

The first sign is broad emotion with no scene: “healing,” “beautiful,” “grateful,” and “life is worth it,” but no place, person, weather, or object. The second sign is perfect structure. Real WeChat posts are rarely polished like a press release.

The third sign is mechanical transitions: “first,” “second,” “in conclusion,” “it is worth mentioning.” The fourth sign is a forced inspirational ending. A coffee photo does not need to become a lesson about destiny.

A rewrite prompt for QClaw

You can send an AI draft to QClaw through WeChat and ask it to rewrite the text. The key is not “make it better.” The key is “make it more like me.”

Rewrite this WeChat Moments post so it sounds more natural and less AI-generated. Rules: 1. Remove phrases such as “in conclusion,” “worth mentioning,” “full of healing,” and forced inspirational lines. 2. Add 1-2 concrete details from the scene, photo, weather, place, or a friend's comment. 3. Make the tone casual, like I am talking to people who know me. 4. Do not turn the ending into a slogan. 5. Keep the original emotion, but make it smaller and more specific. Original: [paste the draft]

Before and after: weekend hiking

AI-sounding version: Today I climbed a mountain and felt the power of persistence. Nature healed me, and I believe every step brings us closer to a better self.

More human version: Halfway up, I seriously considered taking the cable car down. Then the wind at the top made it feel worth it. My legs may file a complaint tomorrow.

The better version does not try to be profound. It has a body feeling, a small conflict, and a joke. That is usually enough.

Before and after: working in a cafe

AI-sounding version: Starting a productive day in the aroma of coffee, becoming a better version of myself through hard work.

More human version: Worked downstairs at the cafe today. Coffee was average, outlet location was perfect. The table next to me discussed renovation for two hours, and somehow I learned half the process.

This version is not “prettier,” but it is more believable. It contains a place, a tiny flaw, and a detail that could only come from that moment.

Five quick prompts to remove AI tone

Daily photo

Make this WeChat Moments caption sound like a normal daily post. Photo: [describe it] My real feeling: [write it] Avoid: big emotions, motivational endings, and perfect structure Keep it under 60 words.

Travel post

Rewrite this travel post so it sounds less like a tourism ad. Place: [place] Visible detail: [weather / food / street / people / small accident] Tone: casual, slightly funny, no brochure language.

Product sharing

Rewrite this product-sharing post like a friend recommendation. What I liked: [specific experience] What is not perfect: [optional] Avoid: hard-sell language, exaggerated results, and fake urgency.

Work update

Rewrite this work update without making it too inspirational. What happened today: [detail] Most annoying or funny moment: [detail] End naturally. Do not add a life lesson.

Shorter and more human

Make this caption shorter and more like a WeChat post. Use shorter sentences. Add one concrete detail. Remove anything that sounds like a brand account. [paste text]

Final human check

Before posting, cover your name and read the caption. If it could be posted by anyone, it needs one more real detail. A small imperfection often helps: bad coffee, windy hair, tired legs, a late train, a friend's weird comment.

AI can save you from a blank page. It should not erase your own trace. The post becomes human when it contains something only you would have noticed.

Use QClaw to rewrite drafts from WeChat

Install the desktop app, bind WeChat, then send your draft and rewrite prompt. Review the final caption before posting.

Check QClaw download notes →