Join Vultr Community
Points, levels & badges
Enter challenges & submit builds
Register to events & meetups
Get your work featured
DocuTrace AI – Evidence-Based Document Review and Verification Assistant
DocuTrace AI is an intelligent document review platform designed to help individuals, educational institutions, HR teams, and organizations analyze, compare, and investigate multiple documents in one place.
Manually reviewing certificates, identity records, employment documents, and other files can be time-consuming and may lead to missed inconsistencies. DocuTrace AI simplifies this process through automated document processing, information extraction, cross-document comparison, and evidence-based investigation.
Users can upload PDF files and images. The platform extracts text from digital PDFs using PyMuPDF and processes scanned documents and images using Tesseract OCR with OpenCV preprocessing. It identifies important information, organizes extracted fields, and compares details across multiple documents to highlight potential inconsistencies such as differences in names, dates, and identification information.
A key feature is Evidence Traceability. Every available finding can be examined through supporting information such as the source document, page reference, extraction method, confidence information, and extracted text. This helps reviewers understand where a finding originated instead of relying on unexplained alerts.
The platform provides a review workflow where reviewers can confirm, dismiss, or mark findings for manual review, add notes, maintain investigation history, explore relationships through a dependency graph, and use what-if simulations to understand related findings. Users can also generate downloadable PDF investigation reports.
DocuTrace AI follows a human-in-the-loop approach. It does not automatically declare documents genuine or fake based on extracted text or mismatches. Potential inconsistencies require human investigation, and official authenticity verification requires confirmation from the relevant issuing authority.
Technology Stack: Python, FastAPI, React, TypeScript, SQLite, Tesseract OCR, PyMuPDF, OpenCV, RapidFuzz, NetworkX, and ReportLab.
Cloud Deployment: The frontend is deployed on Vercel, while the backend API is hosted on Vultr Cloud Compute using Ubuntu, Uvicorn, and Caddy for HTTPS access.
Our goal is to make document review efficient, transparent, explainable, and evidence-driven.
DocuTrace AI – Don't just detect differences. Trace the evidence, understand the findings, and let humans make informed decisions.
Points, levels & badges
Enter challenges & submit builds
Register to events & meetups
Get your work featured