N-CODE AI Studio: An Open-Weight AI Website Audit Side Hustle

Description and use

N‑CODE AI Studio prototype enables freelancers to generate website audit reports combining Lighthouse metrics and open‑weight Gemma 3 recommendations.

Problem

Freelancers and small‑business owners need a low‑cost, transparent website health check that combines measurable Lighthouse findings with actionable AI‑generated advice.

Outcome

The prototype demonstrates that a freelancer can generate a website audit report combining Lighthouse scores and AI‑generated recommendations without relying on proprietary APIs, though the Vercel deployment currently lacks live AI inference.

Workflow

• 1. Deploy Next.js app with Lighthouse metrics endpoint.
• 2. Use Google PageSpeed Insights to fetch lab metrics.
• 3. Run Ollama locally, pull gemma3:1b, and feed metrics to Gemma for recommendations.
• 4. In CI, install Ollama, pull gemma3:1b, run headless Chrome/Lighthouse against target site, and feed results to local Gemma.
• 5. Store Lighthouse JSON and model response as GitHub Actions artifacts.
• 6. Serve audit report via /case-study route.
• 7. Optionally offer human‑led improvements based on report.

Evidence

I built N-CODE AI Studio, an MIT-licensed website auditing prototype for freelancers and small-business owners.

Source

The interface is built with Next.js, TypeScript and React.

Source

The open-weight AI integration uses Ollama and Gemma 3.

Source

The successful October 15, 2026 GitHub Actions run (#38784847640) recorded these single-run mobile Lighthouse lab scores: Performance 57, Accessibility 78, SEO 100, Best Practices 100.

Source

Gemma returned nonempty practical recommendations including review of accessible names, ARIA and page structure, contrast and labeling, and JavaScript/main-thread performance.

Source

An app deployed on Vercel cannot automatically reach Ollama running on my laptop.

Source