Every tool an author pays for, in one place. This is the Author Cockpit: write, plan, check your voice, catch AI slop, research the market and format your book, all in one app. Our Skool is Free to Join. Check out Our Slop Detector Completely Free: http://stopslopstudio.com/detector/ See our list of the TOP FREE TOOLS for Authors: https://stopslopstudio.com/free-tools/ Get the full Cockpit with Stop Slop Premium - Only $20/month for founding members: https://www.skool.com/stopslop

We ran 21 AI models on the same five fiction tasks, cold, no style prompts, and counted every tell in about 79,000 words of their prose. This is the result: which phrases show up, which models lean on them hardest, and the fix for each one. Free for every member. Check your own page against it with the free Detector: https://stopslopstudio.com/detector/ Where these numbers come from. We gave 21 AI models the same five fiction tasks for our 21-model benchmark and kept every word they wrote with no editing and no style instructions (the "cold" run). Then we searched their fiction (the short story, the novel opening and the romance and mystery scenes, about 79,000 words in all) for every tell on our hunt list and counted who used each one, and how often per 10,000 words. Names were counted across all five tasks. Every count below is after we read every raw hit by hand and cut the false positives, like real weight, real flickering lamps, or an ordinary time jump. How to use it. These are words to investigate, not words that are illegal. One hit is a choice. Two on a page, look hard. Three, rewrite. A tell doesn't prove a machine wrote the page. THE BIG THREE (almost every model does these) 1. The em dash. 19 of 21 models used it, 302 times in 79,000 words of fiction. Haiku 4.5 leaned on it hardest (about 130 per 10,000 words), with Muse Wonder next. Fix: swap in a period or a comma, then reread the sentence. Zero is the goal. 2. The default fantasy name. 10 of 21 models reached for the same short list of names: Kaelen, Elara, Silas, Kael, Lyra, Vance and Voss. When a model picks one, it often repeats it as the lead (Gemini 3.1 Pro named its lead Elara and used it 48 times, and Kaelen 32 times; Gemini 3.8 Flash used Kaelen 46 times; Gemini 3.5 Flash-Lite used Silas 48 times; Haiku 4.5 used Kael 34 times). Raw name hits across the run: Kaelen 78, Elara 64, Silas 49, Kael 35, Lyra 21, Vance 20, Voss 18. Fix: if a name arrived first, throw it out. Pull one that fits the culture of your world, and say it out loud.

Welcome to Stop-Slop Studio. I’m Shane. Maybe you’ve been working with AI for a while and you’re tired of getting the same phrases, predictable characters, and prose that sounds suspiciously like everyone else’s. Maybe you have an idea you’ve wanted to make for years, and you’re figuring out where to start. You’re welcome here. We teach you how to get more out of AI while keeping your own judgment and voice in the work. Use it a little or use it a lot. You decide what belongs in your process. Watch the welcome video above, then have a look around: - The Classroom: Start with The First Pass. Bring a page and try the method on something you actually wrote. - The Studio: Explore our writing tools and public benchmark at stopslopstudio.com/studio. You can inspect the work and results for yourself. - Author Cockpit: Our connected creative workspace brings your projects and tools together. Full Cockpit access is part of our paid offering; you don’t need to upgrade to join the conversation and get started here. Writing is our foundation. We’re also exploring how to take a story or world into images, audio, video, and games—and sharing what we learn while we build. Your first step: leave a comment telling us what you want to make and what’s getting in your way. A book? Your first chapter? A world you can’t stop thinking about? Tell us about it. Glad you’re here. Let’s turn imagination into creation, without the slop. — Shane