Cribster

The founders had validated the idea and had no team to build it. Two audiences had to be served by one product: home shoppers who expect a feed that behaves like the apps they already use, and licensed agents who need to see that the platform actually sends them business. Underneath both sits a live MLS feed — authoritative, externally owned, and unforgiving about how you treat it.

• Built the consumer application end to end — swipe-based listing discovery, saved-state and engagement tracking, county and ZIP search with typeahead
• Integrated the regional MLS feed over its RESO Web API, with scheduled sync and status normalization so listing state never contradicts the source of record
• Built the agent side: dashboard, per-listing analytics, contact attribution, agent-to-client invite linking, and a paid premium tier on Stripe
• Stood up the production platform — Postgres, a job queue for bulk work, media CDN, OAuth and magic-link sign-in, rate limiting, staging-first deploys
• Migrated the database off a managed host onto dedicated infrastructure when the vendor's plan was withdrawn, without downtime
• Kept delivery legible: milestones accepted and invoiced in sequence, later work added as written change orders rather than absorbed silently

Cribster is live at cribster.io and in daily use, and the engagement has outlasted its original scope — the founders keep buying more of it, which is the only durable signal that a build was right. They have no engineering headcount and do not need any: one partner owns the application, the data pipeline, and the infrastructure it runs on.