# How to auto-tune your cache lifetime > Stop guessing at a cache lifetime. Adaptive Cache reads your last 24 hours of traffic, recommends a TTL that fits your actual visitor pattern, and applies it in one click. - Category: Cache - Canonical: https://xspeedcache.com/docs/adaptive-cache/ --- Picking a cache lifetime is guesswork. Too short and you rebuild pages that hadn't changed; too long and edits go stale. **Adaptive Cache** removes the guess: it profiles your last 24 hours of real traffic, recommends a lifetime that fits your actual visitor pattern, explains its reasoning in plain English, and applies it with one click. > **Where to find it** > > 1. In your WordPress admin, click **xSpeed Cache** in the left menu. > 2. In the xSpeed Cache sidebar, open the **Cache** group. > 3. Click the **Adaptive Cache** card. > > *Shortcut: open `wp-admin/admin.php?page=xspeed#/cache/adaptive-cache` directly.* > > ⭐ This panel is part of **xSpeed Pro**. ![The 24-hour traffic profile](https://xspeedcache.com/_astro/adaptive-cache-profile.BsLpEahp.png) 1. **Traffic profile (24h)** — what xSpeed sees, with **Analyze** to re-profile. --- ## Settings at a glance This panel has no stored settings — it's an advisor. Everything here is a reading or an action: | Control | What it does | | --- | --- | | **Analyze** | Re-profiles your traffic and produces a fresh recommendation. | | **Apply to Cache** | Writes the recommended TTL into [Page Cache](https://xspeedcache.com/docs/page-cache/). | | **Recommendation** | The suggested lifetime, the delta from current, and the reasoning. | --- ## How it works Click **Analyze** and xSpeed reads the last 24 hours from its hit counter, characterises the pattern, and produces a recommended lifetime with a rationale. **Apply to Cache** then writes that number into [Page Cache](https://xspeedcache.com/docs/page-cache/) — nothing else changes, and your cached pages are never touched. Two safety principles make this trustworthy, and both are worth knowing before you rely on it: 1. **A heuristic computes the number; the AI only explains it.** The recommended hours come from a transparent rule based on your traffic. If you've connected an [AI Provider](https://xspeedcache.com/docs/ai-provider/), it writes the human-readable *reason* — and nothing else. **The AI cannot move the number.** 2. **The result is always clamped to 1–168 hours.** Even a wildly wrong input can't produce an unsafe lifetime. Nothing is applied automatically. A recommendation sits there until you choose to act on it. --- ## Reading the traffic profile The top card shows what xSpeed sees over the last 24 hours: | Metric | Meaning | | --- | --- | | **Character** | A one-word summary of your traffic. This drives the recommendation. | | **Hit ratio** | Share of requests served from cache. | | **Requests / 24h** | Total requests in the window. | | **Current TTL** | Your Page Cache's lifetime right now. | | **Avg / hour** | Mean hourly request count. | | **Peak / hour** | Busiest single hour. | | **Variability (cv)** | How evenly traffic is spread. Around 1 is ordinary; higher means spiky. | **Character** is the one that matters, and it's one of five values: | Character | What it means | | --- | --- | | **Low traffic** | Too little data for a confident recommendation. | | **Cold cache** | Traffic exists but the hit ratio is poor. | | **Spiky** | Bursty — quiet stretches punctuated by heavy hours. | | **Stable, warm** | Steady traffic with a healthy hit ratio. The ideal. | | **Normal** | Everything else. | > 💡 **Variability is the interesting number.** It compares your peak hour against your average. A site averaging 13 requests an hour with a 229-request peak has a cv far above 1 (3.45 in the screenshot above) and is classified **spiky** — and a spiky site wants a different lifetime than a steady one, because its cached pages need to survive the quiet stretches between bursts. --- ## The recommendation ![A heuristic recommendation](https://xspeedcache.com/_astro/adaptive-cache-recommendation.DSHAfCP6.png) 1. **Source label** — whether the rationale came from the heuristic or your AI provider. 2. **Recommended TTL** — with how it compares to your current setting. 3. **Apply to Cache** — writes the number into Page Cache. After **Analyze**, a card appears with the suggested lifetime, the change from your current setting, and the reasoning. For example, on a spiky site: > **24 hours TTL** · No change — *"Traffic is spiky — a moderate TTL keeps peak-hour pages reusable without holding content past its useful life during quiet hours."* If you've connected an [AI Provider](https://xspeedcache.com/docs/ai-provider/), the rationale is AI-written. If not, xSpeed uses a built-in explanation and labels the card **Heuristic recommendation**. Either way **the recommended number is identical** — the AI upgrades the prose, not the maths. "No change" is a perfectly good outcome. It means your current lifetime already suits your traffic. > ⚠️ **The Apply to Cache button is easy to miss.** It sits at the top-right of the recommendation card and renders in a very low-contrast style against the card's pale background — callout **3** above marks where it is. --- ## Related - [How to enable Page Cache](https://xspeedcache.com/docs/page-cache/) - [How to connect an AI provider](https://xspeedcache.com/docs/ai-provider/) - [How to control AI data collection](https://xspeedcache.com/docs/ai-privacy/) - [How to read your cache diagnostics](https://xspeedcache.com/docs/cache-diagnostics/)