What AVMLens does
- Upload a CSV with addresses (city/state/zip optional). Custom identifiers (loan number, parcel ID) pass through to the output for downstream joins.
- Multi-AVM consensus from up to 11 independent sources. The consensus algorithm weights by recency, geographic coverage, and methodology diversity.
- Geo-risk adjustments layer FEMA flood zones, EPA Superfund/TRI proximity, highway/rail noise, and transmission line exposure. The geo score adjusts the consensus value to reflect location-specific risk.
- Batch geocoding resolves addresses at scale (10K+ chunks) with automatic fallback across three providers for coverage.
Output
- Stable CSV schema with consensus value, low/high range, confidence score, geo adjustment percentage, risk-adjusted value, flood zone, and all individual AVM readings.
- Missing fields are blank, not errors. The row still writes. Downstream tools and databases don't break.
- Quality threshold: properties that don't meet minimum AVM coverage get flagged and tokens are refunded automatically.
Tier options
- Tier 1 (2 tokens): AVM consensus only. Fast screening pass to filter on value range.
- Tier 2 (3 tokens): AVM consensus + geo-risk overlays. The standard choice for acquisition screening.
- Tier 3 (4 tokens): Full analysis with comparable sales and aerial context.
Who uses this
- SFR acquisition teams evaluating bulk tapes from brokers or direct sellers.
- Funds running monthly screens across target MSAs to identify mispriced inventory.
- Servicers onboarding new portfolios that need baseline valuations before integration.