Proprietary technology

Vector Sheets Framework

Our proprietary AI framework that transforms how LLMs understand complex spreadsheets — reading logic, not just text.

Powering Model Maestro — the formula-level forensics engine at the core of Diligenz AI™.

Framework capabilities

Eight techniques for reading a spreadsheet as logic.

Structural-Anchor Extraction

Identifies key structural anchors within spreadsheets to maintain layout context while removing redundant data.

Inverted-Index Translation

Creates dictionary-like indexes of non-empty cell contents, significantly reducing token count through intelligent merging.

Data-Format-Aware Aggregation

Groups adjacent cells sharing the same data types or number formats for optimized LLM representation.

Multi-Sheet Cross-Reference

Understands inter-sheet relationships, following formulas and cell references spanning multiple tabs.

Complex Structure Handling

Handles merged cells, pivot tables, and conditional formatting — elements that break competitor solutions.

Large File Token Management

Compression at scale for massive enterprise files (10+ sheets) that would otherwise hit API token limits.

Multi-Format Input

Supports .xls, .xlsx, .csv, PDF, and images via OCR for comprehensive document processing.

State & Security Management

Persistent environment with infrastructure for handling sensitive financial data through sanitization and secure processing.

Powered by VSF™

The logic layer behind Model Maestro.

VSF™ is the technical foundation that allows Model Maestro to treat a financial model as a network of logic rather than a collection of numbers — tracing formula dependencies, detecting structural inconsistencies, and compressing enterprise-scale files without losing fidelity.

Explore Model Maestro →
Begin Due Diligence

See what your current process is missing.

Use our integrated Diligenz Chatbot to ask high-level questions like “What is the primary driver of this quarter’s growth?” — and get an answer based on traced Excel antecedents.