The Ice Beneath

The Challenge

The Ice Beneath is my own project: a dark fantasy trilogy with a wiki deep enough to be worth searching. By mid 2026 the site carried 108 wiki entries and Search Console was telling a blunt story. Pages were being crawled and left unindexed, titles were doubling, duplicate entries were competing with each other, and only 21 of the 108 entries had a meta description. The writing was all there. Nothing about the site let a search engine use it.

The Solution

I audited the site against the crawled-not-indexed data in Search Console and turned the findings into a prioritised 13 item plan. Duplicate entries were merged behind single-hop 301s, doubled titles removed, and descriptions written with bodies expanded across 86 entries in 14 batches. Description coverage went from 21 of 108 to 107 of 108. The search index is now generated from live content at build time by a script I wrote, so it cannot drift from what is actually published, and spoiler content is kept out of both the rendered page and the index. I ran the whole programme with AI agents: the audit, every batch, and the build verification after each one. Underneath it sits a bespoke Next.js front end with a custom MDX engine, interactive maps, and static generation sized for a wiki this interconnected.