How Pagecord uses AI

Pagecord is a one-man band. I use AI to assist with the running of the business. This page documents where and how AI is used.

1. Writing code

I use a coding LLM (currently Claude Code) when I work on the Pagecord app. It's an assistant for writing and reviewing code, finding and fixing bugs, and identifying and fixing security and performance issues.

I review the code myself, the old school way, before it reaches production. Pagecord has a comprehensive test suite that runs before code is shipped. Pagecord's source code is open and free to read on GitHub. Please take a look.

2. Content moderation and spam detection

Pagecord is a user-generated content platform so it attracts spam blogs, and is open to being abused. The T&Cs cover this, and Pagecord has duties under the UK Online Safety Act. To help me identify spam blogs and abuse there are two automated checks, both of which use OpenAI services:

• Content moderation (free and trial accounts only). When a post is published, the title, text and up to five images are sent to OpenAI's content moderation service. This flags whether the content contains banned or illegal categories (violence, porn, hate speech etc).
• Spam blog detection (free and trial accounts only). When a blog is new, and again if it changes during its first month, its title, bio, nav links and up to ten recent posts are checked using a cheap OpenAI model to identify spam blogs (typically SEO link farm stuff).

These checks look at published, public content on free or trial blogs only. Private drafts, deleted posts, email addresses, subscriber lists and account details are not sent. I currently review flagged content manually.

Content sent through these APIs is not used to train or improve OpenAI models. Their terms allow them to keep it for up to 30 days for abuse monitoring, but then it gets deleted. This processing is covered by a data processing agreement that includes the Standard Contractual Clauses. As such, it's in the same UK GDPR framework as all sub-processors listed in the Pagecord Privacy Policy.

Comments, email replies, and contact form messages are checked for spam by CleanTalk. As I understand it, this is reputation-based rather than AI, but I can't say categorically. The same goes for the other third party services Pagecord uses, which are all listed in the Privacy Policy.

What Pagecord doesn't do

Pagecord was built for humans to use the written word to express themselves. This means:

• Customer content is never used to train an AI model.
• Content is never sold or licensed to an AI company.
• There are no AI features in the product.

Every Pagecord blog has a robots.txt to block AI training crawlers and scrapers by default.

AI search indexers are not blocked because they allow your posts to be found and cited. If you want to opt out of this, you can write a custom robots.txt that blocks everything, or disable indexing entirely.

Note that while robots.txt is present, well-behaved companies tend to respect it but other bots and crawlers may not.

I will endeavour to keep this page fresh and accurate. If you have any questions about it, please email [email protected].

Last updated: August 2026