ProEx NIW Letter Platform

Overview

ProEx NIW Letter Platform processes immigration document packages and generates structured EB-2 NIW recommendation letters.

Problem

Manual recommendation-letter drafting from mixed source documents is slow, inconsistent, and hard to scale.

Solution

The platform extracts uploaded content, applies LLM-assisted structuring, builds standardized letter blocks, and exports final deliverables.

Architecture

• backend/: FastAPI services for ingestion, processing, orchestration, and PDF generation.
• frontend/: React UI for uploads, tracking, and downloads.
• docs/: technical and operational documentation.
• attached_assets/: sample/support artifacts.

Tech Stack

• Python 3.11 + FastAPI
• React + TypeScript + Vite
• SQLite (local workflow data)
• PDF tooling (pdfplumber, weasyprint, python-docx)

Setup

1. Backend deps: uv sync
2. Frontend deps: cd frontend && npm install
3. Configure env: cp .env.example .env
4. Run backend: uv run uvicorn backend.app.main:app --reload
5. Run frontend: cd frontend && npm run dev

Results

• Reduced drafting turnaround for NIW package outputs.
• More consistent letter structure across submissions.
• Repeatable export pipeline for client delivery.

Screenshots

Add workflow screenshots under docs/screenshots/ and link here.

Tradeoffs

• Quality depends on source-document quality and extraction accuracy.
• LLM output still requires legal/editorial review before final submission.

Additional Docs

Historical audits and implementation notes are in docs/archive/.