Custom Site Intelligence Audit Tool — Python + Playwright

The problem: Most SEO consultants pay $300/month for a SaaS that runs site audits and produces "actionable insights." The insights are usually a CSV of issues with no severity rating, no business-impact translation, and no competitive context. I wanted a tool I could run on any client site, compare against a competitor, and produce a report that read like an analyst wrote it — not like a crawler vomited findings.

Single-domain audit + competitor comparison, side-by-side. GUI input (tkinter), hit submit, get an HTML report ~90–120s later.

Real Chromium rendering via Playwright. Catches JS-rendered H1s, dynamic meta tags, SPA content, lazy-loaded images, and analytics scripts injected by tag managers — the stuff basic scrapers miss.

Google PageSpeed Insights API integration for mobile + desktop Core Web Vitals on both target and competitor.

Severity-rated insight engine across 8 categories: title/meta, H1 structure, schema, SSL, images, broken links, analytics, performance. Each finding is plain-English with a business-impact translation and a recommended next step.

HTML output that opens in any browser, suitable for client deliverables.

The stack: Python 3.10+ (standard library + Playwright + Google PSI client), tkinter GUI, threading for non-blocking audits, HTML report templating. ~6,250 audits/day max capacity at PSI's free-tier limits.

Why this matters as a portfolio piece: This is the SaaS most consultants in this space pay for. I built my own version because I knew exactly what insights I wanted to surface and what severity model would actually drive client decisions. Productization-ready — V1 roadmap includes multi-pass PSI averaging, LLM-powered conversational insights, batch processing, and HubSpot/Asana field population.

Build philosophy: don't pay subscription fees for tools you can build in a weekend if the cost is your time and the value is shaping the tool to your workflow.