push to main

Showcase · The Circus · 02

Lamplighter

Type any company and see its go-to-market stack. What they run, who they buy from. Fingerprinted live from the edge, then read back to you as a plain-English take.

This page is a walkthrough of that read: the URL that goes in, the fingerprint, and the field read that comes back. A lookup runs in the Worker from the case study.

Cloudflare Workers AI Edge
Read the case study ↗

What it does

Knowing what a company runs, the CRM, the outbound tooling, the analytics, tells you how to sell to them and who they already trust. That picture usually sits behind a paid product. Lamplighter does it from a URL, at no charge, from the edge.

A Worker fetches the company's public surface and fingerprints the go-to-market technology it exposes. Workers AI then writes a field read: a short interpretation of what the stack implies about how they operate, and where an opening might be.

How it works

Inputs

  • A company, given as a URL.
  • The public surface that URL returns. The case study does not describe a login, a database, or a third-party data vendor.

Steps

  • Fetch. The Worker requests the company's public surface.
  • Fingerprint. It names the go-to-market technology that surface exposes: CRM, outbound tooling, analytics.
  • Read. Workers AI takes that list and writes the field read.
  • Return. The response comes back from the edge. One request in, one read back.

The case study describes the build as having no scraper farm and no external infrastructure.

Outputs

  • The list of detected tools.
  • A plain-English GTM interpretation: how this stack suggests they operate, and where an opening might be.

Sample output

Field read

The shape of a response, from the case study. No example company or tool list is published, so this card does not name vendors.

Input
The company URL you typed.
Detected
CRM, outbound tooling, and analytics found on the public page.
Field read
A short note on what that stack implies about how they go to market, and where an opening might be.

Source

Written from the Lamplighter case study. Public code: none under github.com/joelrmartinez as of 19 Oct 2026, so there is no repository link on this page.