You Are The Star Books · Design and build

AI content moderation

Two colleagues in an open-plan office reviewing an AI content moderation dashboard, with messages marked safe, blocked, spam and restricted

For You Are The Star Books, I designed and shipped a Gemini-based multilingual content moderation pipeline using retrieval-augmented generation (RAG) on Google Cloud. It replaced a third-party moderation tool and cut monthly costs from over £8,000 to under £100.

Client
You Are The Star Books
Studio
Rigstore
Role
Design and build
Year
2025 to 2026
Tech
Python, Google Cloud Platform, Gemini, RAG

The challenge

You Are The Star Books relied on a third-party moderation tool that cost over £8,000 a month. Moderation had to work across many languages and keep up as usage grew, without the cost growing with it.

What I did

I designed and shipped a multilingual LLM moderation pipeline built on Gemini, using retrieval-augmented generation (RAG), so the cost follows usage instead of a fixed licence.

  • AI-driven moderation: Gemini detects and flags inappropriate content, misinformation, harmful language and spam.
  • Many languages: moderation works in over 50 languages.
  • Custom moderation levels: each user can set how strict moderation should be for their community.
  • Built to scale: the pipeline runs on Google Cloud with Python and scales automatically with usage.

The outcome

  • The pipeline replaced the third-party tool.
  • Monthly moderation costs fell from over £8,000 to under £100.
  • Moderation now runs in real time across many languages, at a fraction of the previous cost.

Got a project like this?

Tell me what you're working on. I'll reply personally.

Get in touch