Federal fraud detection · public data only
taxbleed.com automatically compares Medicare billing, federal loans, contractor records, sanctions lists, exclusion databases, and H-1B labor filings — then surfaces patterns associated with federal fraud in plain English. Every finding is traceable to a public federal dataset.
—
Federal records indexed
—
Anomalies detected
—
Federal dollars in anomalies
All data sourced from public federal datasets. Anomaly scores reflect statistical deviation, not legal findings.
Three steps. No human judgment until you read the output.
01
Fourteen federal datasets loaded continuously — Medicare billing, PPP loans, EIDL disaster loans, contractor registries, sanctions lists, exclusion databases, corporate ownership, audit findings, DOL H-1B labor condition applications, and USCIS petition data.
67M+ records
02
58 automated fraud pattern detectors cross-reference datasets by stable federal IDs — NPI, UEI, EIN, loan number, LCA case number. Covers Medicare billing fraud, PPP/EIDL loan fraud, federal contracting fraud, sanctions evasion, and H-1B visa wage fraud. Validated at 6x precision lift over random.
58 detectors · 6x validated lift · published methodology
03
Each anomaly gets a plain-English public records analysis — what the two source records show, why the pattern is statistically unusual, which federal laws may be implicated, and three ways to verify independently using public data. Every finding is traceable to a specific row in a public federal dataset.
AI-generated · statute citations · source-traceable
Most detectors fire only on stable federal ID joins — NPI, EIN, UEI, loan number. H-1B detectors use same-row contradictions within a single DOL dataset or exact employer name joins across DOL and USCIS records.
Inflated Payroll
PPP loan exceeds plausible payroll for reported headcount
NPI Laundering
Excluded provider billing under a different provider NPI
Telemedicine Kickback
High Part D cost + large pharma payment from same manufacturer
Pharma Kickback
Extraordinary manufacturer payments correlated with prescribing volume
Excluded Provider Billing
OIG-excluded NPI still generating Medicare Part B claims
Contract Splitting
Multiple awards just below the simplified acquisition threshold
Grant + Material Weakness
Entity received new federal award after repeat audit failures
Address Cluster
Multiple distinct entities operating from one address
Excluded Entity Funded
SAM-excluded company still receiving federal contracts
Peer-Group Drug Cost Outlier
Provider drug spend 5× their specialty's 99th percentile
H-1B Body Shop
H-1B-dependent employer placing 80%+ of workers at third-party sites — non-displacement attestation contradicted by own filing data
H-1B Wage Level Mismatch
Level I (entry-level) prevailing wage certified for a Senior, Lead, Principal, or Architect title — same-row contradiction in DOL LCA data
H-1B Denial Rate Anomaly
Employer with 40%+ USCIS initial denial rate still filing certified LCAs — exploiting DOL's non-adjudication window
All data is sourced from public federal datasets. Every finding can be independently verified by downloading the cited source file from the URL shown in each record.
SAM.gov
Federal exclusions & entity registry
CMS Part B / Part D
Medicare provider billing records
SBA PPP / EIDL
COVID-19 emergency loan data
OIG LEIE
HHS excluded individuals & entities
OFAC SDN
Treasury sanctions list
USAspending.gov
Federal contracts & grants
NPPES
National provider registry
GLEIF
Global corporate ownership chains
Federal Audit Clearinghouse
Audit findings on federal grantees
Open Payments (CMS)
Pharma manufacturer payments to providers
CMS Medicare Revocations
Providers with revoked Medicare enrollment (7,465 records)
DOL LCA Disclosure Data
H-1B labor condition applications — 106,613 certified filings (FY2025)
USCIS H-1B Employer Data Hub
H-1B petition approval and denial rates by employer (FY2023)
DOL WHD Debarment List
Employers barred from the H-1B program by the Wage and Hour Division
Important
Anomaly scores reflect statistical deviation from patterns in public federal data. They are not findings of fraud, wrongdoing, or legal violation. Whether a statistical anomaly constitutes illegal conduct requires investigation beyond what public records can establish. All entity names and data points shown are drawn from public federal government datasets and are accurate as of the date of ingestion.