llms.txt@website/public git:20260827.c6a54e9A

Immutable. This exact content is served forever at /api/v1/blob/6195343277d98128.

# Zero Slop

Zero Slop is an open-source AI writing humanizer and anti-slop checker. It scores the recognizable AI accent in prose and identifies the phrases and structural patterns behind it. It then rewrites the draft and checks that facts were not dropped or invented.

## Core facts

- Website: https://zero-slop.ai/
- Full product reference: https://zero-slop.ai/llms-full.txt
- Source and documentation: https://github.com/manavmishra/ZeroSlop
- Latest release: https://github.com/manavmishra/ZeroSlop/releases/latest
- Current version: 2.6.0
- License: MIT
- Price: Free
- Runtime: Standard-library Python scorer; no third-party packages required
- Privacy: The scorer runs offline and requires no account or server
- Supported environments: Agents that support SKILL.md files, including Codex, Claude Code, Cursor, Gemini CLI, OpenCode, Warp, and Zed

## Install

`npx skills add manavmishra/ZeroSlop --global`

## What makes it different

Zero Slop is not an authorship detector. It is an editorial tool that finds stock phrases, mechanical sentence rhythm, hard-to-read passages, and distracting formatting. Your AI assistant interprets and edits the draft; local tools check facts and protected document structure against the original.

Eight roles move one draft through scoring, interpretation, rewriting, source protection, copy-editing, a read-aloud pass, verification, and a separate fresh-eyes finalizer. Any final polish restarts the last four roles, so the exact text returned to the reader is rechecked. A separate private loop learns only from assistant and final versions supplied by the writer. Shared changes go through a separate release review.

## Examples on the website

The website includes before-and-after sample rewrites for a LinkedIn post, blog introduction, strategy document, X thread, and PowerPoint slide. The displayed writing scores come from the open-source Zero Slop checker. They do not estimate who wrote the text; lower scores mean fewer tracked AI-style patterns.