Which assistants fetch their pages. From their own server logs.
AI crawler traffic, read from your client’s own server logs. Evidence that AI systems requested their pages, not inference.
Trusted by
“Demand and growth has been insane. We’re now building for some of the biggest manufacturers in the country. Search for us on any LLM and you’ll see the results.”
What is AI crawler traffic? In plain words.
AI crawler traffic is the requests AI systems make to a website: to answer someone’s question, to index or train on a page, or to send a visitor. Inciteful reads it from the client’s own server logs and keeps those three kinds of evidence apart, so a crawl is never counted as a visit.
It is direct evidence, from the client’s own servers, that AI systems are reading the pages you ship.
A fetch is not a crawl. A crawl is not a visit.
- Retrieval fetches
- An assistant pulled a page to answer someone.
- Indexing crawls
- A training or indexing bot read it.
- Referrals
- A browser arrived from an assistant.
The pages they fetched, then how to connect a source.
Recently fetched, and verified.
The pages an assistant pulled to answer someone. A fetch is verified when the address it came from belongs to the AI vendor.
- A daily trend, a per-engine table and the recently fetched pages
- Proof uses it: AI crawler visits since a page shipped
Five ways to connect a source.
Send the client’s server logs from wherever the site runs. Health tells you whether events are arriving.
- Awaiting first event, healthy, or stale after 48 hours
- Optional, like every connection: it never blocks a first result
Health reads awaiting first event, healthy, or stale after 48 hours.
From their logs to three clean columns.
- 01
Connect a source
Next.js middleware, a Vercel log drain, a Cloudflare Worker, an Nginx or Apache forwarder, or any server posting with the workspace’s token.
- 02
Watch the health
Awaiting first event, healthy, or stale after 48 hours.
- 03
Keep three kinds apart
Retrieval fetches, indexing crawls and referrals, each in its own column.
- 04
Verify the fetch
A fetch is verified when the address it came from belongs to the AI vendor.
Evidence from their side. Not inference.
- After you ship a page
- Proof shows AI crawler visits since it shipped, when AI traffic is connected.
- For technical clients
- Five ways to connect, so it fits however the site is hosted.
- Optional
- Like every connection, it never blocks a first result.
PlanIn beta.See pricing
Questions about AI traffic, answered plainly.
What is AI crawler traffic?
AI crawler traffic is the requests AI systems make to a website, read from its server logs. Inciteful separates retrieval fetches, where an assistant pulled a page to answer someone, from indexing crawls and from referrals.
How is it different from AI referrals in analytics?
Most analytics tools run in a visitor’s browser, so they only see people who arrive. Server logs also record the requests AI systems make themselves. Inciteful keeps retrieval fetches, indexing crawls and referrals in separate columns.
What does a verified fetch mean?
That the request came from an address that belongs to the AI vendor, so it is not something merely claiming to be that assistant.
How do I connect a site?
Through Next.js middleware, a Vercel log drain, a Cloudflare Worker, an Nginx or Apache forwarder, or any custom server that posts to the ingestion endpoint with the workspace’s token.
How do I know it is working?
The connection shows its health: awaiting first event, healthy, or stale after 48 hours without events.
Is AI traffic required?
No. It is optional and in beta. Without it, Proof reads crawler visits as “not measured” and everything else works as normal.
