How We Perform Operations
We believe that financial and insurance analysis software should be 100% auditable and transparent. This document explains the exact technical architecture, deterministic parsing rules, mathematical verification formulas, and strict consumer safety boundaries governing our utility suite.
1. Zero-Telemetry Client-Side Execution
Unlike traditional insurance lead-generation websites, this platform runs entirely within your web browser. When you upload or paste insurance documents, the data never leaves your device:
PDFs are processed in RAM using pdfjs-dist WebAssembly workers. No file packets are transmitted over HTTP.
We operate zero backend SQL/NoSQL databases, user accounts, tracking cookies, or external analytics endpoints.
Closing your tab or clicking "Reset Session" instantly purges all document strings and extracted figures from memory.
2. The 5-Stage Document Processing Pipeline
Every uploaded declaration page or quote PDF undergoes a 5-stage deterministic evaluation:
The engine streams text items from the PDF DOM, capturing bounding box coordinates, line breaks, and whitespace structure to distinguish coverage tables from payment schedules.
Layout-specific regex patterns identify declarations structures from top insurers including GEICO, Progressive, State Farm, Allstate, USAA, Liberty Mutual, Nationwide, Farmers, Travelers, and American Family.
When declarations contain multiple vehicles, the engine executes a 3-tier cascade (Tier 1: VIN match → Tier 2: YMM alignment → Tier 3: Fuzzy model matching) to ensure vehicle line items are matched accurately.
Carrier jargon (e.g. "BI Liab 250/500", "Comp DED 500") is converted into canonical ISO form definitions (ISO PP 00 01) across Bodily Injury, Property Damage, PIP/MedPay, Collision, Comprehensive, UM/UIM, Towing, and Rental.
Every single extracted numerical value and endorsement is tagged with the exact PDF page number and raw document text citation, allowing policyholders to verify findings with 100% forensic transparency.
3. The Deterministic 4-State Differential Engine
To prevent hallucination, speculation, or subjective bias, our comparison engine assigns every field to one of four objective, factual states:
| State Badge | Trigger Condition | Engine Behavior & Safety Rule |
|---|---|---|
| Same | Identical limits & deductibles across both documents | Records confirmed parity. No changes to coverage scope detected. |
| Different | Variance in numeric liability limit, deductible, or premium | Calculates exact arithmetic delta (±$) and cites both document line items. |
| Not Detected | Coverage line item absent on uploaded summary | Zero-Speculation Rule: The engine never presumes coverage is absent or denied; it explicitly notes non-detection on the supplied summary. |
| Needs Confirmation | Split-limit variance, combined single limit (CSL), or unclear endorsement | Flags line item for manual policyholder review and generates factual questions for their licensed insurance agent. |
4. Mathematical Verification Formulas
All financial audits and calculations use deterministic formulas verified against actuarial and statutory standards:
Measures the percentage difference in annualized or term premium across renewal cycles.
Normalizes multi-year compound price trajectory over policy periods.
Flags discrepancies > $1.00 to expose installment fees or prior-term balance carryovers.
Models cumulative 3-to-5 year statutory surcharge impact vs out-of-pocket repair costs.
5. Consumer Protection Safe Language & Non-Advice Safeguards
Under state Department of Insurance (DOI) standards and editorial integrity principles, insurance is a high-consequence consumer financial domain. Our platform enforces strict boundaries to protect consumers:
The software codebase is audited via automated test fixtures to ensure our system never uses ranking, advising, or disparaging terms such as "recommend", "should", "best", "cheapest", "overcharged", or "error".
We do not sell policies, bind coverage, generate new quotes, or collect lead-generation referral fees from insurance carriers. We are an independent software utility that analyzes documents already in your possession.
A declaration page or quote overview is an informational summary. The actual insurance contract (policy jacket, endorsements, and statutory state filings) legally governs all coverage determinations, exclusions, and claims.
6. Automated PII Masking & Data Redaction
When you export comparison tables or print documents, our client-side sanitization engine automatically masks personally identifiable information (PII) to prevent accidental data exposure:
7. Primary Regulatory Authorities & Standards
Our data structures, state surcharge thresholds, and statutory minimums are validated against primary government and industry authorities:
National Association of Insurance Commissioners standard auto property/casualty model laws and complaint classifications.
Statutory liability minimums (e.g. CA 30/60/15, TX 30/60/25, FL 10/20/10) and official state complaint hotlines.
Federal Trade Commission / CFPB Fair Credit Reporting Act standards for 30-day claims history dispute letters.
8. Revenue Model & Conflict-Free Guarantee
The commercial insurance comparison industry is structurally conflicted: comparison sites make money by selling user contact details to carriers and brokers. Our business model is fundamentally different:
All core document comparison, rate creep tracking, file-vs-pay calculations, CLUE decoding, and total loss discrepancy audits are completely free to use with zero registration.
We will never sell or rent user contact info, policy details, or document text. All data remains in your local browser memory.
Explore Our Privacy-First Comparison Tools
Put our deterministic engine to work on your own documents with 100% local privacy and source-cited precision.