This site is private. Enter the access code to continue.
Incorrect code. Try again.
Everything on this site was built in 72 hours from public data sources and a 30-day API trial. Total cost of data acquisition: $0. This isn't a mockup.
Each step below built on the last. No shortcuts, no synthetic data, no assumptions. Every number on the Sellers and Agents pages comes from this pipeline.
Navigated the Maine Real Estate Commission's license lookup portal. Selected Cumberland County and York County. Downloaded full active agent rosters as CSV exports. Parsed into a SQLite database with license type classification, brokerage clustering, and expiration tracking.
This is Layer 1 of the data pipeline — the credential foundation. Every agent on the scoring table is verified against this database.
Result: 3,730 agents • 249 brokerage offices • 95 Gorham-based agents across 8 offices
Built a custom scraper for the Vision Government Solutions assessor portal (gis.vgsi.com/gorhamme). Reverse-engineered the ASP.NET form submission — the search requires all 10 property type checkboxes and all 44 style checkboxes to return results. Most automated tools miss this and get zero data back. Paginated through the full results set. For each sale, fetched the parcel detail page to capture owner name, assessed value, year built, use code, deed references, and full ownership history.
This is Layer 2 — transaction verification. The assessed values and sale prices shown in the sample Intelligence Report come directly from this data.
Result: 220 verified Gorham sales (2020-2026) • Full parcel details for every record • Ownership chains going back decades
Activated the ATTOM Data Solutions developer portal. 82 API endpoints available on a 30-day free trial. Built an automated property puller that systematically requested AVM Detail for Gorham properties at a paced 1 request/second. Captured automated valuations with confidence scores and value ranges.
This is Layer 3 — property valuation. The AVM estimates on the Sellers page are real ATTOM values, not fabricated examples.
Result: 915 Gorham properties with AVMs • Avg value: $568,285 • Avg confidence: 83/100
Downloaded 350 sold Gorham properties from Redfin with MLS numbers and listing URLs. Manually verified the listing agent and brokerage for 30 curated properties by clicking through to each Redfin detail page. Cross-referenced every agent name against the MREC database — 79% matched to a specific Maine license number, license type, and expiration date.
This is Layer 4 — agent attribution, and the hardest step. Closing this link — from "a house sold" to "this agent sold it" — is what makes the scorecard real. At V2, ATTOM Nexus automates this entirely.
Result: 30 agent-attributed transactions • 24 unique listing agents • 19 confirmed MREC license matches
For each attributed agent, computed the sale-to-assessed ratio across their verified transactions. Cross-referenced license type and brokerage city against MREC. Applied the ratios to a sample property's assessed value to show the dollar impact of agent selection.
The scorecard on the Product: Sellers page uses entirely real data — real agents (anonymized), real ratios, real spreads. The ranking signals on the Agents page describe exactly how these scores are computed.
Result: 41-point spread between best (136%) and worst (95%) performing agents • $205K+ difference applied to a $500K home
Agent credentials, sale prices, property valuations, and listing attribution are all available from public and commercial sources. No brokerage relationship required. No MLS membership needed. The pipeline works.
A working prototype — scraper, database, cross-reference engine, and scorecard — was built in 72 hours by one person. The technology isn't the bottleneck. The data isn't the bottleneck. The only bottleneck is deciding to go.
The 41-point spread between agents isn't a theoretical exercise. It's measured from verified public records. On a $500K home, that spread is $205,000. The seller who chooses Agent A over Agent D walks away with a life-changing difference. That's the product.
Every data source used in this prototype was free (government records) or trial-access (ATTOM). The entire V1 intelligence layer was built without spending a dollar on data. Production ATTOM access adds $100-500/month — funded by the first two agent subscriptions.
Neither could have done it alone.
Every line of code written by AI. Every decision made by a human.
This is what entrepreneurship looks like now.