ClarityClaim

Team Info

Team 15 - kopi-o-kosong

Inspiration

Navigating the Civil and Family Justice Tracking System (CJTS) and managing legal document workflows can be overwhelming, time-consuming, and prone to human error. During legal proceedings, legal teams and self-represented litigants waste hundreds of hours manually reviewing case files, extracting key facts, and formatting tribunal submissions. We were inspired to build CJTS Legal Co-Pilot to bridge the gap between complex legal procedure and modern automation—democratizing legal intelligence and drastically speeding up civil claim processing for the SMU LIT Legal-Tech Hackathon 2026.

What it does

CJTS Legal Co-Pilot is an end-to-end legal assistant built as both a web platform and an integrated Chrome extension. It allows users to: Automate Document Analysis: Instantly analyze, summarize, and extract critical facts, dates, and claims from legal filings and uploaded PDFs. Contextual Co-Pilot Sidebar: Interact with a live AI sidebar while browsing online court portals or legal databases to get real-time procedure guidance.

How we built it

Frontend: Built a responsive, clean user interface (index.html, modern CSS, JavaScript) paired with a Manifest V3 Chrome Extension architecture for sidepanel browser integration.

Backend: Powered by a high-performance Python FastAPI server (main.py) handling asynchronous request routing and document parsing pipelines.

AI Engine & Retrieval: Integrated Large Language Model (LLM) APIs alongside custom prompt engineering to process, summarize, and structure legal text accurately.

Local Development Pipeline: Structured with isolated environment management (venv, PEP 668 compliance) and lightweight local HTTP file-serving for rapid iteration.

Challenges we ran into

Environment Isolation & Deployment: Navigating PEP 668 strict environment management on Linux (Penguin) while orchestrating dependencies (FastAPI, Uvicorn, Pydantic).

Cross-Origin & Extension Communications: Ensuring seamless, low-latency API communication between local file environments (file://), extension sidepanels, and the FastAPI local server (http://127.0.0.1:8000).

Document Extraction & Context Precision: Structuring prompt constraints so the LLM yields accurate, hallucinatory-free legal summaries based strictly on user-uploaded case files.

Accomplishments that we’re proud of

Full-Stack Execution: Successfully bridging a browser extension sidepanel with a robust Python backend in a seamless local dev flow. Instant Legal Extraction: Achieving sub-3-second responses for complex legal document summarization and form pre-filling. User-Centric UX: Designing an interface that abstracts away complex legal jargon into clear, actionable next steps for civil claims.

What we learned

Modern Python Backend Architecture: Deepened our expertise in FastAPI, asynchronous request handling, and CORS configuration for client-side extensions.

Browser Extension Mechanics: Gained hands-on experience with Manifest V3 Chrome Extensions, active tab context sharing, and sidepanel APIs.

Legal-Tech Domain Nuances: Learned how to translate rigid procedural workflows (CJTS rules) into user-friendly automated tools without sacrificing accuracy.

What’s next for Test

Multi-Language Support: Integrating real-time translation for non-English court documents to assist non-native litigants.

RAG Legal Database Vectorization: Implementing Retrieval-Augmented Generation (RAG) over official Singapore court judgments and CJTS practice directions for hyper-accurate case matching.

Automated PDF Annotation: Expanding the Chrome extension to highlight and annotate PDF evidence directly within the browser window.

Built With

• chatgpt-(openai)
• chrome-extension-api-(manifest-v3)
• claude-(anthropic)
• fastapi
• gemini-(google)
• git
• html5/css3
• javascript
• openrouter-api
• pydantic
• python
• uvicorn