Back to Projects

    Title Review Assistant

    Desktop workflow for faster, consistent title opinions

    Completed

    This project laid the groundwork for Cleardeal.

    Windows Desktop Application

    Title Review Assistant is a desktop application built for residential real estate title review drafting. It combines local encumbrance management, optional AI-assisted PDF extraction, and polished Word output with law-firm branding, while keeping lawyer review and approval central to every step.

    Title Review Assistant

    Key Features

    Encumbrance Database Management

    Store, search, edit, and categorize encumbrances with fast lookup during title opinion drafting.

    AI-Powered PDF Analysis

    Extract encumbrances from uploaded title PDFs using OCR and AI, including confidence scoring for each result.

    Human Review Interface

    Review, edit, approve, or reject AI suggestions before anything is committed to the main database.

    Word Document Generation

    Generate professional title review letters with letterhead, sectioned tables, and lawyer signature blocks.

    CSV Import and Export

    Import existing encumbrance datasets and export backups with validation and duplicate handling.

    Firm and Template Settings

    Configure firm identity, letterhead, lawyer defaults, AI provider settings, and optional Word templates.

    Storage and Cleanup Controls

    Persist useful uploaded PDFs and clean rejected files to reduce clutter and keep storage efficient.

    AI Capabilities

    • Supports OpenAI and mock provider modes for testing
    • OCR and extraction flow for scanned title documents
    • Confidence scoring with review-first workflow
    • Effort selector for GPT-5 reasoning levels

    Completed Milestones

    • Core desktop workflow for search, selection, and document generation
    • AI review tab with edit, approve, reject, and confidence display
    • Template-based Word output and lawyer signature block support
    • CSV import/export and in-place database editing workflows
    • File retention and cleanup logic for uploaded PDF analysis

    Problems We Solve

    • Encumbrance analysis is repetitive and slow when done manually from PDFs
    • Maintaining consistency across title opinion letters is difficult without structured templates
    • Searching and curating prior encumbrances can become fragmented across files and team members

    Impact

    • Reduces drafting time by combining search, extraction, review, and document generation in one desktop workflow
    • Improves consistency with reusable encumbrance records and structured section output
    • Keeps legal judgment central through human approval before records are stored or used

    Trust and Compliance

    • Local SQLite storage for firm settings, encumbrances, and AI review records.
    • Configurable AI provider, threshold settings, and optional API key usage.
    • Automatic cleanup for rejected document uploads to control storage use.
    • Clear guidance that human review stays central.

    Technology Stack

    Python
    tkinter / ttk
    SQLite
    python-docx
    PyPDF2 / pdfplumber
    pytesseract + pdf2image + Pillow
    OpenAI (optional)
    Poppler + Tesseract
    DOCX generation

    How It Works

    Search and Upload

    Search existing encumbrances and upload title PDFs when no suitable match is found.

    Review and Curate

    Validate AI results, edit details, and approve selected encumbrances for persistent use.

    Generate

    Generate a polished Word title opinion using your firm letterhead and selected lawyer signature.

    Results You Can Trust

    Designed for legal reliability first: automation assists, but the lawyer remains the final decision-maker.