It is vital to distinguish between a mirror and a scraper.
Recently, there has been a massive uptick in AI companies scraping AO3 to train Large Language Models (LLMs). When fans talk about "stopping mirrors," they are often actually talking about fighting AI scrapers. The OTW has implemented strict robots.txt rules and Cloudflare protections to stop scraping, but a true mirror would require bypassing these protections, further blurring the line between archiving and data theft.
An “AO3 mirror” is an unofficial, often well-intentioned but problematic copy of the Archive of Our Own. While it may provide short-term access in censored regions, it carries substantial risks for users (security, outdated content) and harms creators (loss of control, attribution). AO3’s Terms of Service explicitly forbid mirrors, and the OTW actively enforces against them. ao3 mirror
The recommended approach for accessing AO3 when blocked or slow is to use a VPN, browser extensions, or offline downloads of individual works. For archival purposes, rely on AO3’s own robust infrastructure and the OTW’s commitment to preservation. Unofficial mirrors ultimately undermine the very community and creator rights that AO3 was built to protect.
If you want to mirror an entire subset of AO3 (e.g., every story in a specific tag), tech-savvy users use a command-line tool called wget. However, be extremely careful. It is vital to distinguish between a mirror and a scraper
Mirrors operate in murky territory: copying hosted works can implicate copyright and platform policies. Users and operators should prioritize permissions, takedown mechanisms, and respect for author intent.
An unofficial mirror is not protected by the OTW’s security team. The operator could: Recently, there has been a massive uptick in
The OTW’s legal team is highly protective of the Archive, and for good reason.