case://study-006
AI / ML
Project 006

LeadSniper

LeadSniper is an end-to-end lead generation pipeline for freelancers and small agencies. A local Playwright crawler pulls businesses from Google Places, audits their sites against PageSpeed and a 23-factor scoring rubric, then Gemini 2.5 drafts channel-specific outreach. A Next.js dashboard on Supabase keeps the whole pipeline visible and operable. Splits compute between a local watch-mode scraper (free bandwidth, free RAM) and a hosted dashboard (free Vercel + Supabase tiers).

23
Scoring Factors
Email · WA · LinkedIn
Channels
$0
Infra Cost
AI / ML
Category
The Problem

What Wasn't Working

Lead generation for freelancers is a manual slog: searching directories, visiting dozens of sites, copying emails into a spreadsheet, then writing cold messages from scratch. Most tools solve one slice and cost hundreds a month.

The Solution

How I Fixed It

Split the system into a local scraper (free compute, free bandwidth) and a hosted dashboard (free Vercel + Supabase tiers). The scraper idles in watch mode waiting for search jobs from the dashboard; the dashboard handles pipeline state, scoring, and AI drafts.

Stack

Technologies Used

Next.js 14
TypeScript
Supabase
Playwright
Gemini 2.5
PageSpeed API
Results

Key Outcomes

Runs end-to-end on free infrastructure
23-factor niche-aware opportunity scoring
Channel-specific AI drafts (email · WhatsApp · LinkedIn)
Watch-mode scraper with idle RAM footprint

Want something like this?

Let's build it. I ship fast and I ship clean.