Every AI crawler hit, in one live dashboard.
GPTBot. ClaudeBot. PerplexityBot. Google-Extended. Bytespider. Every hit shows up with timestamp, URL, status and user-agent — the moment it lands.
free trial · live logs from minute one
Watch GPTBot read your pricing page.
The crawler feed updates as hits arrive. Filter by bot, URL or status code. Export to CSV when a stakeholder wants proof.
- 12:04:22GPTBot/pricing200
- 12:04:18ClaudeBot/wordpress-hosting200
- 12:03:55PerplexityBot/docs/backups200
- 12:03:42Google-Extended/200
- 12:03:11
Bytespider/features/woocommerce200
Most hosts hide AI traffic in raw nginx logs.
AI crawlers don't show up in Google Analytics. They don't show up in Plausible. They don't show up in Jetpack. They show up in /var/log/nginx/access.log, mixed in with every favicon request and CSS asset.
By the time you ssh in and grep for ClaudeBot, the conversation is over. The crawler has moved on. You have no idea which pages it read or whether it got 200s or 500s.
Yovale streams crawler hits to the dashboard as they happen. Per-bot panels, per-URL breakdowns, status-code filters — the kind of view a hosting product should have shipped years ago.
Crawler observability, built in.
Four primitives that make AI traffic legible. All four ship with every Yovale plan.
- 0101real time
Live hit stream
Server-sent events push every crawler hit to the dashboard within a second. Page stays open, hits scroll in.
<1s log lag - 0202segmentation
Per-bot panels
Each of the 8 known AI bots gets its own panel — hit count, top URLs, status code distribution, user-agent string.
8 dedicated panels - 0303drill down
Per-URL bot view
Open any URL and see which AI bots have read it, when, and what they got back. Useful for finding 404s only crawlers hit.
URL-level breakdown - 0404ops
CSV export and webhooks
Pipe hits to a webhook for downstream dashboards, or export 90 days of raw events to CSV when an SEO team asks.
webhook + CSV
What real AI bot visibility looks like.
bot tracking stack · how it works
From nginx log line to dashboard panel.
Crawler hits are parsed at the edge, classified by user-agent, deduplicated, and pushed to your dashboard over a server-sent events stream. Nothing to install, nothing to maintain.
Every request hits the nginx access log. A sidecar parser reads the tail, classifies bot user-agents and tags AI crawlers with a canonical name.
8 known AI bots are matched by exact UA string. Unknown bots are kept as 'other AI' until we ship support. False positives are rare.
Classified hits flow to the dashboard via server-sent events. The page stays open, hits scroll in, charts re-render. No polling, no refresh.
Bundled with hosting, not sold separately.
We don't gate bot tracking behind a higher tier. Starter, Growth, Business — all three see every AI crawler hit, all the time.
Starter, 2 sites, full crawler dashboard on both. Annual only.
What pairs with bot tracking.
Goes well with these.
Things buyers ask before trusting the bot dashboard.
Which AI bots are tracked?+
GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, Google-Extended, Meta-ExternalAgent, Bytespider, Amazonbot and Applebot-Extended. Unknown bots are kept under 'other AI' until we add explicit support.
Will tracking slow down my site?+
No. We parse the nginx access log tail in a sidecar process. The web request path is untouched — same nginx, same PHP-FPM, same response times.
Does this work behind Cloudflare?+
Yes. Cloudflare passes the real user-agent through. We classify on the origin so even bots that get cached at the edge still show up when they revalidate.
Can I see hits for a specific URL?+
Yes. Click any URL in the dashboard and you get a chronological list of bot hits, with bot name, status code and response time.
Is there an API?+
Webhook out for new hits, REST endpoint for historical queries. Both authenticated with a per-site token. CSV export covers the no-code case.
How long is data retained?+
90 days, every tier. CSV export available at any point in that window.
Stop guessing whether AI reads your site.
Move a site over in seven minutes. By the time you've made coffee, the crawler dashboard will have its first real data points.