See all your traffic. Humans, bots, and AI agents, told apart.
Browser analytics only sees visitors that run JavaScript. Crawlers don’t, and the bots that do get counted as people. Lume reads every request server-side, on any host or CDN, so you get a clean human count, every bot and AI agent by name, and proof of which ones are real.
Every request identified server-side, measured and named, never blocked.
What your analytics can’t see.
Browser analytics, a CDN’s bot dashboard and Lume each see a different slice of your traffic. Lume sorts every request into 17 agent types against a catalog of 120 agents, on whatever host you run.
| Browser analytics | CDN bot dashboard | Lume | |
|---|---|---|---|
| Crawlers and bots that never run JavaScript | Browser analytics Invisible | CDN bot dashboard Yes | Lume Yes |
| A human count with the bots taken out | Browser analytics Bots that run JS count as people | CDN bot dashboard Requests, not people | Lume Yes |
| AI assistants as their own referral channel | Browser analytics Manual channel setup | CDN bot dashboard No | Lume Built in |
| Verified vs. self-declared crawlers | Browser analytics No | CDN bot dashboard On its own edge | Lume Yes |
| Any host, any CDN, or several at once | Browser analytics Yes | CDN bot dashboard Its own network only | Lume Yes |
| The verdict in your own code and reports | Browser analytics No | CDN bot dashboard Inside its own rules | Lume API and export |
The segments in your traffic, and what to do about each.
Most analytics lumps these together or never sees them. Lume tells them apart, so each one comes with a decision you can actually make.
Humans
Real people browsing your site. The clean count everything else is measured against.
What you do: Report conversion and engagement on people, not crawlers.
AI-referred humans
People who clicked through from ChatGPT, Perplexity, Gemini and other AI answers.
What you do: See which assistant sends them to which pages, and invest there.
AI agents
Software acting for a person, like an assistant reading a page or comparing products.
What you do: Make sure they can reach what they came for instead of filtering them out.
AI crawlers
Bots that fetch your pages to train AI models or ground their answers.
What you do: Catch the 403s and 404s they hit before you drop out of AI answers.
Search & utility bots
Search crawlers, uptime monitors, feed readers and SEO tools.
What you do: Keep them out of your human numbers and spot wasted crawl.
Spoofers / unverified
Traffic using a well-known crawler’s name that fails the checks for it.
What you do: Hand the verdict to your WAF or CDN rules.
Two jobs, one source of truth.
Some teams want bots under control. Others want to grow the traffic AI is starting to send. Both need to know who is actually on the site.
For sysadmins, security and engineering
Keep bots in check
- Is that really Googlebot?
- Every request claiming a well-known crawler is checked, so a copied user-agent gets flagged.
- Who is eating my bandwidth?
- Bots and AI crawlers ranked by requests, operator and page, without grepping a log.
- Are my rules blocking the right things?
- Base WAF and CDN rules on verified identity and stop challenging real customers and good crawlers.
For SEO and marketing teams
Grow from AI search
- Can AI crawlers reach my pages?
- See which crawlers fetch which URLs and the status codes they get back.
- Is AI sending anyone?
- AI-referred visitors by assistant and landing page, kept apart from search and Direct.
- Can I trust my numbers?
- Sessions, bounce and conversion measured on humans, with the bots removed.
Verified, not just self-declared.
A user-agent string is just a claim, and anyone can copy it. The big search and AI operators publish ways to check their crawlers, and Lume runs those checks on every request that uses their name: signatures, reverse DNS and published IP ranges. It works the same on any host or CDN, and the verdict is yours, in the dashboard or through the Identification API.
Verified vs. self-declared. Know which is which.
From first request to full picture.
Connect
Point Lume at your CDN logs or edge, or add one SDK call. No client JavaScript.
Classify
Every request sorted as human, bot, or AI agent, across 17 agent types.
Verify
Signatures, reverse DNS, and IP ranges checked for the 7 operators that support it.
Act
See the split live and unsampled, down to operator and page, and fix what needs fixing.
The questions people ask first.
My CDN or Google Analytics already shows this. Why add Lume?
Browser analytics can’t see crawlers, because they don’t run JavaScript, and it counts the bots that do as people. A CDN sees every request, but only on its own network, and its bot data stays in its own dashboard. Lume reads requests server-side on any host or mix of CDNs, tells humans, bots, and AI agents apart in one place, and hands you the result through export and an API.
Verification only labels the honest bots. What about the rest?
True, and that’s why it isn’t the whole product. Verification catches anything using the name of one of the 7 major operators that publish a way to check their crawlers, so a copied “Googlebot” gets flagged. Everything else is still classified and shown, and anything that fails a check is labeled unverified rather than hidden. A pass tells you who an agent is, not that it behaves; the analytics show what it actually did.
Does Lume block bots?
No. Lume measures and verifies; your WAF, CDN or app decides what to allow, challenge or block. Basing those rules on verified identity is how you stop catching real customers and good crawlers by mistake.
Will Lume get my site cited in AI answers?
No, and nothing can promise that. Lume shows whether AI crawlers can reach your pages and whether AI answers send people back. What you do with that is up to you.
Do I have to add code to my site?
No. Connect your CDN logs or an edge integration and nothing touches your pages: no client JavaScript and no added latency. A TypeScript SDK and a REST endpoint are there if you’d rather send events yourself.
Start free. Scale when your traffic does.
Starter
- 5M events/mo
- 3 projects
- CSV export
- Identification API
Paid plans pause at their included allowance unless you opt into overage, billed per whole million events and never above your spend cap. See full rates on pricing.
Latest from the blog.
301 redirects: how to use them effectively and handle edge cases
A practical guide to 301 redirects: when to use them over 302/307/308, how link equity passes, and how to handle chains, migrations, and edge cases.
How to tell if Cloudflare or your firewall is blocking AI crawlers
A diagnostic flow for finding which layer stops GPTBot, ClaudeBot, or OAI-SearchBot: robots.txt, Cloudflare's AI crawler settings, a WAF rule, rate limiting, or your origin. Plus when blocking them is the right call.
Crawl budget: what it is and how to protect it
Crawl budget is the number of URLs bots crawl on your site in a set window. Learn what wastes it, how logs expose the leaks, and why AI crawlers now spend it too.
How to detect AI agents like ChatGPT agent on your website
AI agents browse your site in a real browser on a person's behalf, so they look nothing like crawlers. Here is the verification ladder that tells signed, declared and hidden agents apart.
Know your traffic. All of it.
Connect your CDN or add one SDK call. Free plan, no card required.
Start for free →Privacy-first by design: analytics that never stores a full IP, truncated to a network prefix before storage. US-hosted. Hashed tokens only.