ahmedak/defluff

Overview

A Python library that deterministically detects AI-generated prose by analyzing textual features without using machine learning models or API keys. It operates entirely offline, relying on rule-based heuristics to flag likely slop.

Best for

Best for
Developers needing a quick, offline, zero-cost slop check with no reliance on third-party services.

Use cases

• Flag AI-generated content in code comments or documentation
• Filter out synthetic prose from scraped datasets
• Add a lightweight slop check to CI pipelines

Notes

A Python library that deterministically detects AI-generated prose by analyzing textual features without using machine learning models or API keys. It operates entirely offline, relying on rule-based heuristics to flag likely slop.

0 stars on GitHub. Last updated 2026-06-28. Licensed MIT.

Use cases

• Flag AI-generated content in code comments or documentation
• Filter out synthetic prose from scraped datasets
• Add a lightweight slop check to CI pipelines

Pros

• No external dependencies or API costs
• Fully deterministic and reproducible results
• Simple to integrate as a pure Python library

Cons

• Limited accuracy compared to model-based detectors
• Very low adoption (0 GitHub stars) with minimal community validation
• May produce false positives on intentionally stylized prose

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.