Patreon stops asking AI bots not to scrape — and starts blocking them

Patreon is no longer relying on polite requests to keep AI scrapers off creator pages. The company is working with Cloudflare to block bots that train AI models on creators’ content without permission. The change moves Patreon from asking scrapers to stay away to actively stopping unauthorized AI training traffic.

The technical shift is from robots.txt-style notices to network-level bot blocking. Robots.txt only tells well-behaved crawlers what they should not fetch. It does not stop scrapers that ignore those rules. By pairing with Cloudflare, Patreon can filter and block training-oriented bot traffic before that content is pulled at scale for model training.

For engineers and builders, this is a practical signal that content platforms are treating AI training scrapes as an access-control problem, not a courtesy problem. If you ship product features that depend on open web crawl access, or if you run a site that hosts user-generated content, assume that “please don’t scrape” is no longer considered a sufficient control. Permission and enforcement are becoming part of the same product surface.

In market terms, Patreon is aligning creator-platform defense with a broader industry move away from voluntary crawl etiquette. Creators monetize original work on Patreon; unlicensed training on that work undercuts that model. Cloudflare’s role places this inside the same bot-management stack many large sites already use for abuse and automated traffic, rather than treating AI scrapers as a separate, soft policy issue.

The practical takeaway is to treat unauthorized AI training crawls as something to detect and block, not only document. Watch how platforms define permitted access for model training, how bot blocks are applied at the edge, and whether creators get clearer controls over who can use their work for AI systems. For product and infrastructure teams, the next design question is which crawlers are allowed by policy and which are rejected in the request path.