The platform

Natural-language intelligence over your hardest data

Ultra-Data turns plain-English questions into trustworthy answers — conversation, tables, charts, and correlation maps — over your own data and the public record, without raw PII ever leaving your on-prem infrastructure.

Ask in plain English

Pose a natural-language question and get a narrated, well-visualized answer across hundreds of millions of business records — no query language to learn.

Five-store hybrid retrieval

Structured filters, full-text, vector similarity, geospatial, and graph relationships are combined into one coherent answer in seconds.

Private by architecture

Records stay on-prem behind a four-layer redaction gate. Only sanitized snippets ever reach the cloud model — raw PII never leaves your infrastructure.

Reasoning you can audit

A planner-critic-narrator pipeline plans each retrieval, checks its own work, and cites every claim to source records — with a downloadable, cryptographically signed evidence bundle.

400M+ validated records

Query a proprietary, validated dataset of over 400 million records in plain English — the curated foundation the rest of the platform builds on.

Bring your own data

Upload your own lists and embed them directly into the ask path — your records and ours, answered in one query. Keep them current with scheduled auto-enrichment, and pay only when a row matches.

Public-record enrichment

Consolidate public information from across the web — property, business filings, court, and civic & financial records — into one cited profile. Every finding carries a confidence band, never a precise score, and links back to its source.

Privacy & compliance, built in

Consent provenance on every record, a self-serve privacy opt-out, and a built-in data-removal path for data subjects. Sensitive fields are redacted before anything reaches the cloud model.

Findings from property records, business filings, court records, and death records are public-record information — probabilistic, not verified, and presented with a confidence band, never a precise score. They may not be used, in whole or in part, for tenant, employment, credit, or insurance eligibility decisions, or for any other purpose covered by the Fair Credit Reporting Act.

What it does

Five ways to work the record

Powerful search

Ask in plain English across a proprietary, validated dataset of over 400 million records plus the public-record source set — property, business filings, court records, civic and financial filings — and get a narrated, cited answer in seconds. No query language to learn.

Advanced analytics

A cross-entity link and pattern graph surfaces direct connections — co-occupants, shared addresses, shared business officers, owner-to-asset, related parties — alongside key connectors, shortest paths, clusters, and ring or cycle detection. Every connection carries a confidence band, never a precise number.

Dynamic collection

A continuously updated collection: each panel shows when its source was last updated, change-over-history tracks how findings move across refreshes, and opt-in profile monitors alert you when watched facts change. Findings are kept current on each source's own refresh cadence.

Deep research & due diligence

The platform assembles a structured due-diligence dossier on an entity — Executive Summary, Identity Matrix, Asset Profile, Litigation, Professional Verification, and Sources & Methodology — with every finding stated as a confidence band and each section carrying an explicit coverage status. It is research material, not a consumer report.

Person, place or thing

Entity-anchored profiles span person, business, property, and asset, so a single anchor resolves across the records that reference it. Start from a person, a place, or a thing and follow the connections outward.

Change log — refresh-cadence-bound, not real-time. Each entry is dated as of its source refresh.

Audiences

Who is it for

Ultra-Data is built for research, due diligence, legal, and government work — never for tenant, employment, or other FCRA-covered screening.

Law firms & legal

Litigation support, asset tracing, and due diligence backed by cited public-record findings.

State & local government

Investigations, fraud and recovery work, and program integrity over public records.

Federal government

Research, oversight, and analysis across the public-record and federal-filing landscape.

Corporate legal

Counterparty diligence, conflicts review, and entity research for in-house teams.

Corporate & academic research

Market, policy, and academic research over validated and public-record data at scale.

Nonprofits

Investigative journalism, watchdog, and public-interest research grounded in the public record.

Transparency

Sources & coverage

Coverage is honest about its edges. Some sources are nationwide; others cover only specific states today and are labeled as such.

National coverage

  • Property records — nationwide county coverage
  • Court records
  • Death records — pre-2014 public records
  • FEC campaign finance
  • U.S. Census
  • USAspending federal awards
  • IRS Form 990 nonprofit filings
  • FAA aircraft — national registry
  • NPPES national provider identifiers

State-specific coverage

  • Business filings — SEC/EDGAR (national) plus Secretary-of-State registries in Florida, New York, and California
  • UCC secured filings (liens) — California and Connecticut
  • Professional & occupational licenses — Virginia

Coverage varies by source and by state. Per-state sources (such as UCC secured filings and professional & occupational licenses) reflect only the states currently covered and are not nationwide. All findings are public-record information presented with a confidence band, never a precise score, and may not be used for tenant, employment, credit, or insurance eligibility decisions, or for any other purpose covered by the Fair Credit Reporting Act.