Pakistan Justice Navigator

Pakistan Justice Navigator is an AI-assisted legal understanding and case-preparation product for Pakistani legal contexts. It helps people explain a situation in their own words, organise facts, explore potentially relevant legal context, work with documents, and prepare better questions for a qualified legal professional.

This project provides informational and preparation assistance. It is not legal representation, professional legal advice, or a guarantee of any legal outcome.

Repository structure

frontend/                  Next.js citizen experience and legal research UI
backend/                   FastAPI triage API
frontend/services/legal-rag/  Official-source research service
frontend/data/              Corpus and member manifests
frontend/supabase/          Database migrations and seed data
docs/                      Product, API, setup, and verification documentation

Frontend quick start

Requirements:

• Node.js
• npm
• Python for the corpus checks used by the project scripts

cd frontend
npm install
npm run dev

Open http://localhost:3000.

Useful commands:

npm run typecheck
npm run build
npm run test:core
npm run check

Optional services

The frontend can use the following services when configured:

• BACKEND_API_URL — triage API
• LEGAL_RAG_API_URL — official legal research service
• NEXT_PUBLIC_SUPABASE_URL and NEXT_PUBLIC_SUPABASE_ANON_KEY — browser-safe Supabase values

Keep LLM credentials and Supabase service-role keys server-side. Never put secrets in NEXT_PUBLIC_* variables.

Backend triage API

cd backend
py -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8001

The API documentation is available at http://localhost:8001/docs.

Legal research service

The official-source research foundation lives in frontend/services/legal-rag. Its setup, corpus acquisition, provenance rules, and ingestion workflow are documented in:

• Member 4 setup
• Master implementation plan
• Team commands

Product experience

The landing page uses a warm editorial legal design system:

• Scroll-linked hero and product storytelling
• Bronze legal-object imagery with transparent cutouts
• A globe identity visual in the LEGAL / AGENT transition
• A pinned video story with five scroll-controlled parchment cards
• Responsive desktop and mobile layouts
• Reduced-motion support
• Accessible semantic sections, navigation, focus states, and image descriptions

Runtime media is stored in frontend/public:

frontend/public/references/
  books-gavel.webp
  fist-scales.webp
  legal-agent-globe.webp
  story-card.webp
frontend/public/video/
  legal-story.mp4

Legal-information safety principles

The product should:

• Say “helps you understand”, “supports research”, and “potentially relevant”
• Preserve the distinction between user-provided facts, AI organisation, source candidates, and verified evidence
• Keep source identity visible where legal material is presented
• Encourage professional review for important decisions

The product must not claim guaranteed accuracy, guaranteed outcomes, official government advice, or that it replaces a lawyer.

Documentation

The integrated implementation documentation is available in frontend/README.md. It includes the current corpus state, Supabase migrations, environment boundaries, and the full verification workflow.