Skip to content
Live — Shipped and running in production2026

Full-stack developer

Three readings
Client
A real estate agent enters an address and, in under 45 seconds, gets a report that defends the listing price to the seller. Not a CMA, not an appraisal — a pricing defensibility system.
Recruiter
Full-stack on an AI SaaS for US real estate agents: a fifteen-module OpenAI pipeline, PDF rendering through PDFMonkey templates, Stripe subscriptions across three tiers, SendGrid email and Chart.js dashboards.
Engineer
One AI run, two renders: the agent's full analysis and a sanitised client version are built from the same result, so the two can never disagree about the numbers.

What it does

  • A pricing report from a single address in under 45 seconds
  • Fifteen analysis modules: market response, pricing and valuation, price-path scenarios, a seller stress test
  • A client-ready PDF for the listing appointment, plus the agent's full internal version
  • A confidence level that tells the agent exactly what would raise it
  • Three subscription tiers, from five reports a month to unlimited

Under the hood

  • Each report snapshots the agent's tier when it is created, so a plan change mid-cycle never rewrites a finished report
  • The client render is stripped of the internal analysis before it is generated, never hidden with styling
  • Report limits are metered per billing cycle against the tier, straight from the Stripe subscription

Built with

  • OpenAI
  • PDFMonkey
  • Handlebars templates
  • Stripe subscriptions
  • SendGrid
  • Chart.js

Signal path

architecture
compstieragentclientMarket datacomps · scenariosAgententers an addressStripetier · limitsAI pipeline15 modulesReportone run · <45sPDFMonkeyHandlebarsAgent PDFfull analysisClient PDFsanitised