Quirio
AI assistant for Texas residential real-estate agents covering listing search, comparative-market analysis, client notes, and document drafts. A small group tried it during early access. I archived it in April 2026 after failing to find a repeatable way to reach users.
What it was
An AI assistant for residential real-estate agents in Texas. It combined listing search, comparative-market analysis, client notes, and document drafting in one interface backed by MLS data.
What we built
Listing discovery over the live Texas residential MLS — natural-language queries against the structured and unstructured fields ("three-bed under 600k in 78704 with a yard and not on a busy street"). A CMA generator that pulled sold comps weighted by recency, similarity, and proximity, and produced a draft narrative the agent edited rather than wrote. A lightweight conversational CRM ("leads I haven't followed up with in 10 days"). And a document drafter (offer letters, listing descriptions, follow-up emails) where Quirio drafted and the agent edited.
The final product used a Next.js and TypeScript front end with an Express and TypeScript backend, model APIs for the conversational layer, and Vercel for hosting. It worked with real MLS data under the licensing constraints that implies.
What I learned
A small early-access group tried the product and supplied feature-level feedback. That was useful, but it was not enough evidence to claim retention, willingness to pay, or product-market fit.
Building against live data taught me to treat model output and third-party services as unreliable inputs: validate structured output, retry bounded failures, log decisions, and keep deterministic checks around consequential actions.
I also learned the operational basics of shipping and maintaining a product: deployments, regressions, support, billing, and data-licensing edge cases.
What broke
Distribution. Real-estate agents are already served by an entrenched vendor ecosystem (MLS providers, CRM platforms, transaction-management tools) with incumbent contracts and trade relationships I had no way to compete with. Cold outbound was slow. I underestimated this by a multiple.
Market access. I started in 2024, when prospective users already depended on MLS, CRM, and transaction-management vendors with existing contracts and workflows. I underestimated how difficult it would be to enter that stack.
Focus. Quirio plus trading-algo plus high school. The right call was to commit to one or wind Quirio down sooner. I split attention longer than I should have, and the marketing function suffered first.
Lessons
I built the product and didn't crack outbound. Two different jobs; I treated them as one.
By 2026 the moat is data access or distribution, not the model. Vibe-coded "AI assistant for X" is now a category that anyone can ship in a weekend.
I notified the early-access group, took the product offline, and archived the code instead of leaving an unsupported service running.
Model output is untrusted input. Validation and hard constraints matter more than how fluent the interface feels.
Status
Archived April 2026. The landing page is still up at quirio.io; the product behind it is off. Source open under Apache 2.0: frontend (Next.js + TS), backend (Express + TS).
mahimn · quirio · post-mortem · apr 2026