PolicyLensAI — Intelligent Policy Analysis Platform

PolicyLensAI is an AI-enabled web application that helps policy researchers analyze, compare, and improve policy documents. It integrates document processing, evidence retrieval, and large-language-model (LLM) analysis into an interactive research workflow. Built with modern RAG services, the app supports structured policy review, automated excerpt extraction, and live drafting assistance.

---

🧭 Core Features

• Policy Upload & Extraction Upload policy PDFs or Word files. The system automatically extracts text and structures it into analyzable sections.

• AI-Driven Analysis Uses OpenAI’s GPT-4o model to identify key policy excerpts, align them to codebook frameworks, and generate improvement recommendations.

• Evidence Integration Automatically derives keywords and queries Google Custom Search to retrieve academic and grey literature supporting or challenging policy content.

• Framework Comparison Annotates aligned and misaligned policy segments, highlighting excerpts directly in the rendered PDF, with hover-based pop-up explanations.

• Interactive Research Tools Create and edit projects, codebooks, codes, subcodes, and excerpts. Collaborate with other users, manage media files, and view structured comparisons.

• Live Writing Assistant A “Live Writing” mode suggests evidence-based sentences and structures in real time while drafting policy briefs, frameworks, or strategic plans.

• Data Management Includes migration and schema scripts for updating database structures, sample data generation utilities, and dummy excerpt populators.

---

⚙️ Tech Stack

• Backend: Flask, SQLAlchemy, Flask-Migrate
• Frontend: Jinja2, Bootstrap, JavaScript
• Database: SQLite — supports PostgreSQL migration
• AI Services: OpenAI GPT-4o, Google Custom Search API
• Embedding/Indexing: Sentence-Transformers (MiniLM-L6-v2), FAISS
• Environment: Python 3.10+ with .env for API keys

---

🚀 Quick Start

1. Clone & Setup

   git clone https://github.com/<your-username>/policylensai.git
   cd policylensai
   python3 -m venv venv && source venv/bin/activate
   pip install -r requirements.txt

2. Configure Environment Edit .env:

   OPENAI_API_KEY=<your_openai_key>
   GOOGLE_API_KEY=<your_google_key>
   GOOGLE_CSE_ID=<your_cse_id>
   

3. Initialize Database

   python init_db.py
   python sample_data.py

4. Run the App

   python run.py

   Then open http://localhost:5000

---

🧩 Key Modules

• app.py — Main Flask application and route definitions
• services/analysis.py — Handles LLM-based policy–framework comparison
• services/policy_analyzer.py — Evidence retrieval and improvement generation
• models/models.py — ORM schema for users, projects, media, and excerpts
• forms.py — Flask-WTF form definitions for uploads, profiles, and projects
• sample_data.py — Generates realistic sample content for testing

---

🧠 Authors

Developed under the PolicyLensAI initiative at the African Population and Health Research Center (APHRC) Lead AI Expert: Dr. Tatenda Duncan Kavu, Lead Full Stack Developer: David Wanambwa, Policy Expert: Henry Owoko Odero

---

📜 License

© 2025 APHRC. All rights reserved. For research and non-commercial use only.