mikiarlo3/ai-copywriter

What it solves

AI-generated copy often suffers from two main problems: it is either too generic and fails to capture attention (e.g., using words like "unlock" or "revolutionize"), or it sounds like a chatbot, leaving obvious markers of AI writing. This project provides a portable agent skill that combines high-conversion copywriting techniques with a rigorous "humanizing" process to create text that reads like it was written by a person and earns attention.

How it works

The skill operates as a set of instructions (a SKILL.md file) that can be loaded into any LLM or agent harness. It uses a two-pronged approach:

1. Strategic Copywriting: Instead of focusing on product features, it focuses on the reader's emotional state and simplicity. It interviews the user to gather a specific Ideal Customer Profile (ICP), a category, and a true story with concrete numbers to avoid generic output. It then produces multiple variants of copy (headlines, microcopy, LinkedIn posts, etc.) based on the reader's feeling.
2. Humanizer Engine: It applies 33 detectable AI writing patterns (based on Wikipedia's "Signs of AI writing" guide) to audit and rewrite text. This includes removing "significance inflation," AI-specific vocabulary, and formulaic structures to ensure the result is natural and human-like.

Who it’s for

Marketers, founders, and writers who use AI to draft copy but want to avoid the "AI smell" and create high-converting, specific, and human-sounding marketing materials.

Highlights

• Humanizer Audit: A rigorous 33-point check to strip out AI-isms like "testament to," "landscape," and excessive em-dashes.
• Interview-Driven Process: Proactively asks for real numbers and stories rather than inventing facts or using generic adjectives.
• Multi-Format Support: Specialized templates for clickbait titles, microcopy, subject lines, LinkedIn posts, and strategic blog posts.
• Portable Format: Delivered as a Markdown file that works across ChatGPT, Claude, Manus, and other LLM interfaces.
• Voice Calibration: Ability to analyze a user's own writing samples to match their personal rhythm and quirks.