"Every time I asked an AI tool to write something, it added a long dash and the medium dash. The em dash. This one ( - ). And the en dash. This one (-). At first, I let it slide. But the more I wrote, the more it felt out of place. It did not sound like me. It felt like a tiny glitch in my voice. Then I realized I was not alone."
"Because it doesn't always work. AI models can be like rebellious teenagers. You tell them not to do something, and they do it twice to prove a point. Some ignore the rule, some swap in en dashes pretending they're being "helpful, " and some just keep adding them because... honestly, I have no idea why. On top of that, I also use other tools (besides ChatGPT) that don't always have default instruction settings, like a small local model on my laptop, Gemini, or Claude."
Dashy Drop targets the subtle punctuation patterns that mark AI-generated text, specifically em and en dashes. The tool post-processes output from any writing workflow to remove or normalize those dashes, restoring punctuation that matches a human voice. Prompt-based instructions proved unreliable because models sometimes ignore or substitute rules. The tool works universally across different models and interfaces, including local models and cloud services, by cleaning text after generation rather than trying to control models' behavior. The goal is cleaner, more polished output that preserves human tone while retaining AI assistance for drafting and ideation.
Read at Medium
Unable to calculate read time
Collection
[
|
...
]