Not everyone wants to integrate an API. Sometimes you just need a documented property valuation - with confidence intervals, model metrics, and comparables - without writing code.
That's why we built the automated valuation model (AVM) dashboard. Search for any Canadian property, click a button, and get a comprehensive valuation report. No coding required.
How it works
- Log into your Houski account
- Go to the AVM dashboard
- Search for any property by address using the autocomplete
- Select your property from the results
- Choose your payment method:
- $5 via Subscription - if you have an active API subscription, the charge is added to your monthly bill
- $5 One-time - pay directly via credit card with no subscription required
- Access your AVM report immediately
Your purchased AVMs live in your dashboard forever. Unlike static appraisals, Houski AVMs update automatically. Check back next month or next year and valuations reflect current market conditions.
Track your property values over time
Your dashboard shows all your purchased AVMs in a table with:
- Address - links to the full property page
- Sale Price (initial) - estimated sale price when you purchased
- Sale Price (current) - today's estimated value
- Sale Price Change - gain or loss since purchase, highlighted green or red
- View Report - link to your full AVM report
Monitor property values across your portfolio and see how markets affected each property since you valued it.
What's in an AVM report
Each report includes three independent valuations:
- Estimated list price - likely listing price today
- Estimated sale price - likely closing price
- Estimated monthly rent - rental market value
You also get confidence scores, model accuracy metrics, and up to 10 comparables within 5km. The same data that powers our API, via a simple web interface.
Reports are shareable - send the link to clients or colleagues without requiring a Houski account.
Who needs a documented AVM?
The AVM dashboard is for professionals who need more than a price estimate. Each report includes confidence intervals, model accuracy metrics (R-squared scores, median absolute percentage error, hit rates), and comparables - documentation required for lending decisions, insurance underwriting, and investment analysis.
Useful for:
- Lenders assessing collateral with documented confidence levels
- Insurers determining coverage with supporting comparables
- Investors performing due diligence and tracking portfolio values
- Appraisers supplementing traditional methods with data-driven estimates
- Legal professionals needing documented valuations for disputes or estates
- Property managers monitoring asset values across properties
To track your own property's value for free (no AVM purchase), use the My Properties dashboard. The paid AVM is for when you need the full report with confidence metrics and comparables.
Two ways to pay
- One-time payment - $5 USD per AVM, charged immediately. No subscription required.
- Subscription billing - With an active API subscription, AVM purchases are added to your monthly bill. Same price, consolidated billing.
Both give lifetime access with automatic value updates.
When to use the API instead
The dashboard works for one-off valuations or small portfolios. To value hundreds or thousands of properties, use the AVM API.
The API returns the same data programmatically, letting you integrate valuations into your systems and process bulk requests.
Both cost the same: $5 per property. The difference is just how you access the data.
The most accurate valuations in Canada
You get the same models in the dashboard and the API. Current per-city accuracy metrics, including median absolute percentage error, R-squared, and the share of predictions within 10% of actual value, are published live on the AVM information page. A low median error means half of valuations land within that percentage of actual value on held-out test data. For a $1,000,000 home with a 4% median error, that is within $40,000.
We have city-specific models for many Canadian cities, with province-wide and country-wide models covering everywhere else. The AVM information page lists the current set of city models alongside their accuracy.
How we measure accuracy
We publish detailed accuracy metrics for every model on our AVM information page:
-
Median absolute percentage error (MAPE) - How accurate typical predictions are. A 5% median error means half our estimates are within 5% of actual value. For a $500,000 home, within $25,000. Median rather than mean avoids skew from outliers.
-
R-squared Score - How much price variation our model explains. An R-squared of 0.96 means the model accounts for 96% of differences. For context, city name alone might get R-squared around 0.30. Square footage, bedrooms, and neighborhood might get 0.65. Our models use far more features effectively. Current per-city R-squared values are on the AVM information page.
-
Within 10% - The percentage of predictions within 10% of actual. At 80%, 4 of 5 estimates for a $500,000 property fall between $450,000 and $550,000.
Hierarchical model selection
Models train at three geographic levels:
- City models - Most accurate, trained on city-specific patterns
- Province models - When no city model exists (all provinces covered)
- Country models - Baseline for areas without more specific models
We automatically use the most specific model available, ensuring best accuracy while providing estimates for every property in Canada.
View live accuracy metrics
The AVM information page shows current metrics across Canada:
- Accuracy rankings for every city model
- Province comparisons
- Sale price, list price, and rental metrics separately
- Detailed performance statistics for each model
- Exactly how reliable your valuations are for any location
We publish these because when you're making decisions about property values, you deserve to know how reliable your data is.
Get started
If you have a Houski account, head to your AVM dashboard.
If you're new, create an account - it's free. You only pay when you purchase AVMs.
For developers integrating valuations into applications, see the AVM API documentation.
