Skip to content
ProjectsAvailable

Verified Bots

Which crawlers are real, and how to prove it.

verifiedbots.devFree & self-hosted

Three ways to know it is really them.

A user agent is a claim, not proof — anything can send one. The tier is how much of that claim a bot's operator has actually made checkable.

Tier 1

Fully verifiable

The operator publishes a machine-readable IP feed, or supports Web Bot Auth. You can confirm a request came from them without trusting anything it told you.

Tier 2

Verifiable

No live feed, but a documented reverse-DNS pattern, ASN or static range. Checkable, with a round trip and a little more care.

Tier 3

Listed only

The identity is known and the user agent is documented, but nobody published a way to check it. Recorded honestly as unverifiable rather than quietly implied to be safe.

Of the 204 bots catalogued, 110 carry a recipe you can actually run — 57 at tier one and 53 at tier two. The directory says which, including when the answer is that there is not one.

How it works

Five steps, in the order they happen.

  1. 01

    One bot, one file

    Every entry is a YAML file validated against a published JSON schema. Adding a bot or correcting a range is a pull request, not a private database edit.

  2. 02

    Sources refresh daily

    Operators publish their own ranges, and the pipeline pulls them every day rather than trusting a list someone wrote down once.

  3. 03

    A bad feed cannot poison it

    CIDR sanity checks, a diff alarm when a range list swings unexpectedly, and automatic fallback to the last known-good feed when a source looks wrong.

  4. 04

    Every bot gets a tier

    Tier 1 is fully verifiable — a machine-readable IP feed or Web Bot Auth. Tier 2 is verifiable by reverse DNS, ASN or static ranges. Tier 3 is listed only: the identity is known, but nobody published a way to check it.

  5. 05

    Then it is just files

    It builds to all.json, per-category and per-tier sets, individual bot profiles, user-agent patterns, and plain .ips lists a firewall can read directly.

The facts

Everything below is verifiable. Nothing is rounded up.

Bots catalogued
204 across 10 categories
With a verification recipe
110 of 204
IP ranges tracked
9,227
Largest categories
Search engines 49, Monitoring 34, SEO tools 33
AI crawlers and assistants
31
Refresh
Daily, with poisoned-feed safeguards
Formats
JSON, .ips lists, RSS
Licence
MIT for code, CC-BY-4.0 for data
Affiliation
Independent — not affiliated with Cloudflare
Figures as of
1 August 2026