How Onsite works, in detail.

Data pipeline

Onsite runs a continuous ingestion pipeline across 4,000+ municipal permit sources (Socrata, ArcGIS REST, assessor APIs, and county portals). Every permit is normalized into a single schema before it reaches the scoring layer.

• Sources
  Government databases in all 50 states

  City permit portals, county assessor records, state licensing boards. We track new APIs and retire stale ones automatically.

• Refresh
  Continuous sync, hourly retry

  High-volume sources refresh hourly. Lower-volume sources run every 6 to 24 hours. Typical lag from filing to platform: under 24 hours.

• Normalization
  Address, ZIP, geocoding, deduplication

  Every record is geocoded, tagged to a ZIP, and deduplicated against existing permits. Inconsistent fields are repaired via our geo_repair service.

• QA
  Source reconciliation & trust scoring

  Broken or drifting sources are quarantined. Trust is tracked per source so downstream scoring can weight more reliable feeds.

Territory model

Access is scoped to ZIP codes, not a vague region. When you subscribe, you reserve exact ZIP codes up to your plan’s cap, and lead access is enforced at query time on every request.

Lead scoring

Every lead gets a score from 0 to 100 computed at ingestion time. The score combines permit valuation, recency, type, and location against a calibrated model that classifies each lead as Hot, Warm, Medium, or Cold.

• Signals
  Valuation, recency, type, density

  Higher permit values, newer filings, and project types matching your trade contribute positively. Leads in saturated ZIPs get a density penalty.

• Temperature
  Hot / Warm / Medium / Cold

  Buckets are calibrated against the distribution across the whole database. Today: 49% Hot, 10% Warm, 10% Medium, 31% Cold.

• Recomputation
  Nightly drift correction

  Scores and days_old fields are recalculated each night so stale leads fall in priority automatically.

Owner enrichment

Raw permit data rarely includes contact details. Onsite runs an enrichment pipeline across county assessor records, parcel databases, and verified third-party sources to attach owner names, phones, and emails to each lead.

• Primary
  County assessor & parcel records

  Owner name and mailing address come from authoritative public records. Coverage is strongest in metros; we are continuously expanding rural counties.

• Contact
  Phone & email enrichment

  Verified contact details are attached where available. No per-lead unlocks. No surprise charges.

• Accuracy
  Quality labels per field

  Every enriched field ships with a quality marker so you know when something is verified, partial, or minimal.

Compliance & licensing

Onsite is built for licensed contractors only. Access requires a valid contractor license number and state, and lead usage is audited through the built-in pipeline tracker.

• Licensing
  License number required at onboarding

  Collected during signup and verified against state licensing boards where APIs are available.

• Data use
  Territory-scoped, audited

  You only see and export leads inside your reserved ZIPs. Any attempt to access outside territory is blocked at the API layer.

Security & billing

Security is baked in end-to-end. Authentication uses JWT with HMAC-SHA256 signing, payments run through Stripe (PCI DSS Level 1), and every request is rate-limited per plan tier.