austin-weeks/miasma

Trap AI web scrapers in an endless poison pit.

What it solves

It prevents AI companies from scraping public websites for training data by actively punishing scrapers rather than just blocking them.

How it works

Miasma sends "poisoned" training data and self-referential links to scrapers. This creates a loop of low-quality content (slop) that is designed to degrade the quality of the AI models being trained on it.

Who it’s for

Website owners who want to protect their content from being used as AI training data without incurring a high compute burden.

Highlights

  • High performance with a minimal memory footprint.
  • Available as a Docker image, Rust library, or pre-built binary.
  • Designed to degrade model quality through poisoned data loops.

Related

  • Project
    itsOwen/CyberScraper-2077

    An AI-powered web scraping tool that uses LLMs to intelligently extract and structure data from websites while providing features to bypass bot detection and navigate paginated content.

  • Project
    h4ckf0r0day/obscura

    A lightweight, high-performance headless browser engine written in Rust, designed as a drop-in replacement for Puppeteer and Playwright to power AI agents and web scrapers with minimal resource usage.

  • Project
    spider-rs/spider

    Spider is a Rust library for fast, concurrent web crawling and scraping. It streams pages as they’re fetched, only launches headless Chrome for JavaScript‑heavy pages, and can run locally or on the managed Spider Cloud service (which adds proxies, stealth, and scaling). The crate provides a simple API, configurable limits, and bindings for Node, Python, and a CLI, making it useful for feeding web data into AI/RAG pipelines, SEO monitoring, or any large‑scale scraping task.

  • Project
    dondai44423/donsetch

    A high-performance Rust-based web research tool for AI agents that provides stealthy fetching, keyless search, and crawling without requiring API keys.