Guide
Clean up your Ghost tags across the whole blog
Tags drift. Over a few years and a few authors, you end up with “AI” and “artificial-intelligence” and “a.i.” all in use, near-duplicates that fragment your topic pages, one-off tags that tag nothing else, and inconsistent capitalization. The Ghost admin lets you rename or merge tags one at a time, which is fine for a small fix and miserable for a real cleanup. When the job is normalizing taxonomy across the whole archive, you want to see and edit all the tags at once — which means a recipe over the whole connected blog.
Tags travel with each post
This is squarely in Specter’s wheelhouse, because tags are one of the fields it syncs. Specter connects to your Ghost blog in the browser and brings every post into one workspace with its tags intact; edits flow both ways, and publishing updates the tags on Ghost. So a taxonomy cleanup becomes ordinary editing across the whole archive, with the full picture in front of you instead of buried one post at a time. There’s nothing to install.
Survey first, then apply the merges
For the messier job of figuring out what the taxonomy should be, run an AI recipe — and make the useful first step a survey, not an edit:
Read the tags across every post. List all distinct tags with how many posts use each, and group the ones that are obvious duplicates or near-synonyms. Don’t change anything yet — just propose a normalized tag list and the merges to get there.
Review that proposed mapping — this is where your judgment about your own topics matters — then have it apply the merges you approve:
Apply this tag mapping to every post: [your approved mapping]. Only change tags. Leave each post’s body, title, and other metadata untouched.
Browsing your archive and reviewing the proposal are free; only the AI run spends credits, and your workspace opens with 500 free.
Review the diff, then publish
Tag changes ripple into your topic pages and navigation, so review the diff carefully — Specter shows exactly which posts would update before anything goes live, and reviewing costs nothing. Every publish keeps a snapshot behind it, so a mapping you misjudged is one revert away from undone.
The same approach powers the rest of the bulk jobs here — SEO edits, meta descriptions, and internal links. Tags are just another field, and once your whole blog is in one workspace, fixing them everywhere at once stops being a chore you avoid. Subscribe now →
Prefer a plain find-and-replace over .md files on your own machine — handy when you already know exactly which tags map to what? The desktop and open-source edition gives you the folder to do it locally.