Learn from auction outcomes.
The training approach pairs historical final winning bids with property and auction data available at the bid snapshot. No hindsight-only information.
HARRIS COUNTY, TX / EARLY ACCESS
Predict the
winning bid.
Foreclosure auction bid prediction is the mission. Machine learning built on historical final bids, designed to help you understand the likely winning price before you bid.
Final-bid data Three quantiles Harris County first
Get early access.
Built for iPhone Safari and Chrome. No app download.
PROPERTY IN FOCUS
A suited investor requests property intelligence on a phone. The camera moves through the phone screen into a grayscale Houston aerial view, rapidly approaches a home, and resolves into a digital model rotating at 45 degrees above horizontal. Leader labels show the demo's living area of 3,450 square feet, year built 1965, and lot size of 18,000 square feet. Illustrative values: P50 bid $290,800, 2025 tax appraisal $412,000, comp $489,000, rehab $127,500, and potential profit $70,700 before other costs. Pause or choose a scene using the controls. No real drone feed or property scan.
01 / THE WORKSPACE
Your next bid.
In your pocket.
Start with the predicted bid. Then make the numbers yours. Adjust bid price, comp estimate, and rehab allowance to see the effect on equity and profit.
- 01Start with the prediction.
Compare P10, P50, and P90 bid estimates.
- 02Check the property.
Review the facts and prior sale years.
- 03Account for repairs.
Explore the rehab range and its limits.
SAMPLE DATA / INTERACTIVE DEMO
Illustrative scenarios, not measured returns. No account needed.
PROPERTY BRIEF
DEMO MODEIllustrative data. Not a live valuation or an instruction to bid.
Demo account & connection
DEMO // AUTH & BILLING ARE SIMULATED
THE MACHINE-LEARNING APPROACH
Not an asking price.
The auction's final bid.
Our prediction target is what the winning bidder pays, not a listing price or tax appraisal. Property and rehab context support that core prediction.
See more than one outcome.
CatBoost is our primary model family, with separate P10, P50, and P90 estimates. XGBoost and LightGBM are challengers, not a stacked model.
Prediction meets judgment.
Use the bid range alongside editable comp and rehab assumptions. Scenarios do not alter model predictions. No estimate guarantees a winning bid or a profit.
02 / THE PERSON BEHIND THE PRODUCT
Meet the founder.
A focused tool.
An accountable builder.
Leonardo Ruelas Jr.
Founder / Trustee Sniper
Trustee Sniper is being built around a clear goal: make the final winning foreclosure bid more predictable. Property intelligence and rehab scenarios give that prediction context, in one practical mobile-web workflow. Harris County first.
Meet Leonardo on LinkedIn (opens in a new tab)03 / BEFORE YOU JOIN
A few straight answers.
Is the demo using live auction data?
No. It uses illustrative fixtures, not live model scoring or verified property valuations. Sample profits demonstrate the math, not actual or promised returns. The demo is not an instruction to bid.
What happens when I join the waitlist?
We will email you when access opens. Joining requires only an email address, not an account or payment. No launch date or subscription price has been announced.
Do I need to download an app?
No. Trustee Sniper is being built for the browser, with iPhone Safari and Chrome as the priority. Harris County is the initial service area.
YOUR NEXT MOVE
Be ready for what comes next.
Join the Harris County early-access waitlist.