How to Run a Freshness Audit and Prioritize What to Update
Audit your published library for stale content over Content Hub MCP and receive a prioritized list of what to update first.
A freshness audit measures the time since each post was published or last updated. It flags posts older than a threshold, typically 180 days, as stale. Published content loses accuracy over time, and large libraries often accumulate stale posts unnoticed.
Connect Claude Desktop to Infrasity's Content Hub over MCP. You can then audit the whole library in one message and receive a stale and fresh count plus a prioritized list of what to update first.
Before you start
- Topics are published with real URLs and recorded publish dates, because freshness is calculated from post age.
- The library has meaningful history. An audit of a few weeks of posts is of limited value.
Fetch the freshness report and analysis
Request the freshness report for the blogs of the Infrasity organization, together with an analysis.
Claude opens with a summary: 88 published posts audited against a 180-day threshold, of which 50 are stale (57%) and 38 are fresh (43%). It then lists the oldest posts to refresh, each with its exact age and URL. Examples include "Organic vs Paid Marketing on Reddit for SaaS" at 384 days and "Checklist for GEO and AEO in 2025" at 378 days. The full stale list spans publish dates from August 2025 to February 2026, and Claude offers to export it.
The analysis reports four findings:
- More than half of the library is stale. Stale content affects traditional SEO, because Google favors updated content in competitive results. It also affects AI-search visibility, because LLM crawlers weight recency when selecting citations.
- Duplicate entries inflate the stale count. Some titles appear twice under identical or near-identical URLs. Examples are the "Mintlify vs. GitBook" comparison under two URLs and a B2B content marketing piece published three days apart under the same URL. Claude identifies these as likely duplicate Content Hub records and recommends a data cleanup before relying on the freshness figures.
- GEO/AEO/LLM-visibility content dominates the stale cohort. This is the strongest-performing topic cluster overall. Titles such as "Checklist for GEO and AEO," "How to Rank on ChatGPT," and "AEO vs GEO" are aging into stale territory. Model capabilities and ranking factors change monthly in this area, so outdated advice damages credibility.
- Freshness improves for content published after roughly February 2026. Recent sprint output is current. The problem is a backlog of older content.
Request takeaways
Request a condensed list of takeaways for the analysis.
Claude returns a numbered list, each item paired with supporting data:
- 50 of 88 posts have not been updated in six or more months.
- The GEO/AEO/LLM-visibility cluster is the best-performing content on the site and is aging out fastest. Refresh it first.
- Duplicate topic records inflate the numbers and warrant a cleanup.
- Staleness is a legacy problem. Content from recent sprints is current.
- The oldest 10 to 15 posts (350 or more days) need attention first, because their statistics and screenshots are most likely outdated.
- A recurring refresh cadence is required. A one-time cleanup does not prevent 57% of the library from becoming stale again within six months.
A single freshness request now produces a prioritized refresh list in one Claude Desktop conversation. Each recommendation traces to a specific post and its age.
To audit your own blog, open app.infrasity.com and run a freshness audit.
