Guide
Audit your blog for things that have gone stale
Old posts quietly go wrong. A statistic from three years ago is now misleading, a “currently” that is no longer current, a price that changed, a tool you recommended that shut down, a “this year” that points at the wrong year. None of it is dramatic, but it accumulates, and it erodes trust with exactly the readers who arrive from search and assume the post is maintained. Finding all of it by re-reading your archive is the kind of task nobody ever gets to. An AI recipe across your whole connected blog does.
Audit first, edit second
The instinct is to ask AI to “update my old posts,” but that invites it to rewrite things that were fine. The better pattern is to separate finding from fixing. Specter connects to your Ghost blog in the browser and brings every post into one workspace, so a recipe can read all of them at once — and you start by asking it to flag, not change:
Read every post. Produce a report of likely-outdated content: specific statistics or numbers that may have aged, dates and “this year”/“currently” phrasings, named tools or products that may have changed or shut down, and any claim that reads as time-sensitive. For each, give the post, the exact line, and why it’s a flag. Change nothing.
That gives you a worklist instead of a black-box rewrite. Crucially, an AI may not know the current fact — its knowledge has a cutoff and it can be confidently wrong — so treat the report as “things to verify,” not “things that are settled.” You or a quick web check confirms the actual current value before anything changes.
Apply the fixes you trust
Once you’ve verified the real numbers, you can make the corrections directly, or hand the confirmed list back to a recipe for the mechanical edits:
Apply these specific corrections to the posts named: [your verified list of old value → new value]. Change only those facts and adjust surrounding wording only as needed to read naturally. Don’t touch anything else.
Review the diff, then publish
Read the diff — factual edits especially deserve a second look, since this is your credibility on the line — then Specter confirms which posts would update before going live. Reviewing is free, and every publish keeps a snapshot behind it. Connect Specter via the first-sync guide if you haven’t yet; your workspace opens with 500 free credits and only the AI run spends them.
A fact-freshness audit pairs naturally with refreshing old posts at scale: one keeps the content competitive, the other keeps it correct. Both are only practical because the whole blog is now a workspace your tools can read end to end.
This is a sensitive area in one respect worth stating plainly: AI is good at spotting what might be stale and bad at deciding what is true now. Keep yourself in the loop on the facts, and this becomes one of the highest-trust maintenance passes you can run. Subscribe now →
Prefer to run the audit on your own machine, over .md files with your own AI? The desktop and open-source edition does it locally.