Live, internal use

Squint

An audit-driven pitch generator for agencies: it turns a website audit into a client-winning, competitor-benchmarked report in under two minutes.

My role
Positioning → build
Status
v1.0 · Live internally
Stack
Chrome MV3 + Node + Gemini
Segments
Agencies · Individuals
The 30-second version

Live and in daily internal use. Discovery reframed a crowded dev QA tool into a pitch generator for agencies: a website audit in, a competitor-benchmarked, client-ready report out in under two minutes. Next milestone is the hosted web app.

Squint report screenshot

The problem

Generic pitches get ignored

Every agency, CRO consultant, and freelancer faces the same cold-outreach problem. Manual audits take 4–8 hours per prospect. Nobody does them at outreach scale, so cold emails stay generic and convert below 1%. Individual site owners have the opposite problem: they know their site underperforms but can't articulate why, and can't afford a $5K agency audit to find out.

  • Existing tools (Lighthouse, axe, WAVE, PageSpeed) are built for engineers: raw diagnostics, never the business question.
  • Nobody answers: "what is this costing me, and compared to whom?"
  • The original positioning framed this as a dev QA tool: a crowded, free, commoditized space. The real asset is the persuasive output artifact.

Discovery

Two segments, one growth loop

AG

Agencies & consultants · Revenue

CRO shops, web design agencies, freelance UX consultants (1–50 seats). JTBD: "Make my cold outreach impossible to ignore, and justify my retainer."

Willingness to pay: high, $79–$249/mo
IN

Individuals · Funnel

E-commerce owners, indie founders. JTBD: "How bad is my site really, and what do I fix first?"

Free reports carry Squint branding → market to agencies

Backlog

User stories that shaped the build

US-01 As an agency owner, I want an evidence-backed audit of a prospect's site generated in minutes so my cold outreach doesn't sound generic. Must
US-02 As an agency owner, I want the report benchmarked against the prospect's strongest competitor so the pitch quantifies exactly what they're losing. Must
US-03 As an individual site owner, I want a free branded report so I understand what's wrong with my site without paying for an audit. Must
US-04 As an agency owner, I want white-labeled reports so I can send them under my own brand. Should
US-05 As an agency owner, I want engagement tracking on sent reports so I know which prospects actually opened them. Should
US-06 As an agency, I want a bulk-audit mode so I can prospect at scale instead of one site at a time. Could

Requirements

Spec'd as testable acceptance criteria

RequirementPriority
AI competitor discovery scoped to the prospect's category, with a 7-day cache per domainMust
Dual full-page capture with stealth mode and a scroll-and-stitch fallbackMust
Telemetry harvest: axe-core, Core Web Vitals, trackers, tech stack, contrast, with zero canned fallback contentMust
Hosted, zero-friction web app: remove the localhost/extension dependency blocking non-technical usersMust
White-label PDF branding and bulk audit runs for agency accountsShould
AI-generated "after" mockup of the prospect's page with top issues fixedCould

My role

Repositioning the product, then building the pipeline

Squint originally shipped as a developer accessibility/QA extension: a crowded, free, commoditized space. I repositioned it around the actual asset: the persuasive, branded report that wins business, not the diagnostics that produce it. From there I defined the two-segment growth loop, wrote the capability ladder toward a defensible product, and built the audit pipeline (competitor discovery, capture, telemetry harvest, and AI critique) end to end.

What I built

An audit pipeline that ends in a pitch, not a diagnostic

AI competitor discovery

Gemini identifies the prospect's strongest direct competitor from page metadata, with manual override and preset benchmarks.

Dual full-page capture

Server-side stealth Puppeteer with splash-screen detection, backed by an in-browser scroll-and-stitch fallback.

Telemetry harvest engine

axe-core accessibility, Core Web Vitals, page weight, tracker/pixel detection, tech-stack fingerprinting, contrast inspection.

Gemini multimodal critique

Screenshots + DOM + metrics produce a UX alignment score, executive summary, and CRO recommendations, delivered as a dark-themed PDF and interactive HTML report.

Under the hood

URL in, pitch-ready report out

1Prospect URL submitted → Gemini identifies their strongest competitor (cached 7 days)
2Both sites captured full-page via stealth Puppeteer (or scroll-and-stitch fallback)
3Telemetry harvested from both: accessibility, CWV, trackers, tech stack, contrast
4Gemini critiques screenshots + DOM + metrics side-by-side → score, summary, recommendations
5Report delivered as branded PDF (email or download) + interactive HTML link

Outcome

Live and in daily use: the wedge is now distribution

The full audit-to-pitch loop (competitor discovery, capture, telemetry, AI critique, branded delivery) is built, live, and in daily internal use as a Chrome extension + local backend. The repositioning from "dev QA tool" to "pitch generator" reframed the entire go-to-market. The next milestone is the hosted web app that removes the extension/localhost barrier standing between agencies and the product.

Roadmap

The 10/10 ladder

Next

Hosted, zero-friction web app: paste a URL, get a report in under 90 seconds. No extension, no keys.

Then

Evidence-grade claims pinned to annotated screenshots, plus an AI "after" mockup: the launch-moment differentiator.

Later

Proprietary benchmark corpus: percentile claims that compound with every audit run.

More case studies

Want a discovery sprint or a prototype like this one for your product? Book a call.