Individual Research Project · 2025 / 2026

AI-Powered Civil Law Assistance Platform for Sri Lankan Citizens

LegalEase LK — Empowering everyday citizens with accessible, bilingual legal guidance powered by AI

Vimansa D. A. S  ·  IT21346668

Project ID25-26J-553
InstitutionSLIIT
DegreeBSc (Hons) IT
SupervisorMs. Uthpala Samarakoon
Year2025 / 2026
Overview

Project Abstract

LegalEase LK is a legal assistance platform powered by AI, aimed at enhancing access to legal information for everyday citizens in Sri Lanka. Current platforms like LankaLAW and LawMate mainly cater to legal practitioners and lack straightforward, multilingual legal clarifications appropriate for the average person. This study seeks to address this issue by creating an accessible platform that delivers legal advice in Sinhala and English via a chatbot powered by a Large Language Model (LLM), as well as a lawyer referral system that assists users in locating suitable legal experts based on specialization and geography. The backend utilizes Node.js/Express, while the frontend employs React, guaranteeing scalability, performance, and user-friendliness. The system relies on selected legal data sources like "Edineda Jiwithayata Neethiya" to produce precise and comprehensible answers. The anticipated result is heightening legal awareness, better accessibility to lawyers, and strengthening public confidence in the justice system.
The Problem

Inaccessible Legal Information

Most Sri Lankan citizens struggle to understand their legal rights due to complex language and platforms designed only for legal professionals, leaving the public dependent on unreliable advice.

The Solution

AI-Powered Legal Chatbot

A bilingual (Sinhala/English) conversational AI chatbot using a custom fine-tuned LLaMA 3 model trained on Sri Lankan civil law datasets for plain-language legal explanations.

Key Impact

Smart Lawyer Matching

A recommendation system that connects citizens with certified lawyers by specialization and geographic district — making legal help accessible even in rural areas.

Civil Law Domains & Technology Stack
Domains Covered
Property Law Vehicles & Highways Marriage & Divorce
Technology Stack
LLaMA 3 8B React.js Node.js / Express MongoDB Atlas Supabase Unsloth
Domain

Research Domain

Literature survey, research gap, problem statement, objectives, and methodology.

Background & Literature Survey

Access to legal information in Sri Lanka has long been limited for ordinary citizens who struggle to interpret legal documents and understand their rights. Although digital legal platforms exist — such as LankaLAW and LawMate — they are primarily designed for lawyers, researchers, and law students, and do not offer multilingual support or plain-language guidance for the general public.

Internationally, AI has been adopted in legal domains through platforms such as DoNotPay, LawDroid, and LegalZoom. However, these systems are built around Western legal frameworks and do not apply to Sri Lankan law or local languages. LegalEase LK bridges this gap using LLM-based conversational models, information retrieval systems, and modern web frameworks like React and Node.js.

Research Gap

Research Problem

Ordinary Sri Lankan citizens struggle to access and understand legal information due to complex legal language and professional-focused platforms. Existing digital legal systems are designed mainly for lawyers and students and do not provide simple, multilingual explanations or conversational support. As a result, people lack legal awareness, depend on unreliable advice, and face difficulties finding suitable lawyers, especially in rural areas. This highlights the need for an accessible, AI-powered legal assistance platform tailored for the general public.

Research Objectives

Main Objective: To design and develop an AI-powered, multilingual legal assistance platform that provides Sri Lankan citizens with simple, accessible legal information and helps them connect with suitable legal professionals.

Methodology

  1. User enters a legal question through the React web interface in Sinhala or English.
  2. The question is securely sent to the Node.js/Express backend.
  3. The custom fine-tuned LLaMA 3 8B model processes the question using vector similarity on the legal knowledge base.
  4. The model identifies the law category and retrieves the most relevant legal sections from MongoDB.
  5. The platform uses the identified category to fetch suitable lawyers from the lawyer database.
  6. A clear legal explanation and best-matched lawyers are displayed to the user bilingually.

Technologies Used

AI / ML
LLaMA 3 8B (Unsloth fine-tuning), GGUF format, vector embeddings, sentence transformers, Retrieval-Augmented Generation (RAG)
Frontend
React.js, HTML, CSS, JavaScript, Bootstrap / Material-UI
Backend
Node.js, Express.js, REST API, ngrok API tunnel, Notification service
Database
MongoDB Atlas (lawyers + qa_chunks collections), Supabase PostgreSQL
Milestones

Project Milestones

All project assessments with details, dates, and marks allocated.

AssessmentDescriptionDateMarksStatus
Project ProposalProblem statement, literature survey, objectives, and methodology defined.Sep 202510%Completed
Progress Presentation ILLM model training initiated; dataset preparation complete; system design finalized.Nov 202515%Completed
Progress Presentation IIChatbot integration, lawyer recommendation module, and frontend development progress.Jan 202620%Completed
Final AssessmentComplete system demonstration with evaluation results and final documentation.Apr 202640%Upcoming
VivaIndividual oral examination evaluating research depth and technical understanding.May 202615%Pending
Development Phases
Phase 1 — Foundation
Requirement analysis, dataset preparation (Edineda Jiwithayata Neethiya Parts 1–3), system architecture design
Phase 2 — Development
LLM fine-tuning with Unsloth, backend API, React frontend, chatbot integration, lawyer recommendation module
Phase 3 — Evaluation
Functional testing, user acceptance testing, performance evaluation, cloud deployment, final documentation
Documents

Project Documents

All produced documents and pending submissions.

Completed
Project Charter
Project initiation and scope — 1 document
View →
Project Proposal Report
Full research proposal — 25-26J-553 · IT21346668 · Sep 2025
View →
Check List Documents
Assessment checklists and supervisor review forms
View →
Pending
Final Research Document
Main thesis report — not yet submitted
Pending
Presentations

Presentation Slides

Slides from past presentations and provisions for upcoming ones.

PP — 01
Project Proposal Presentation
Presented: December 2025
View Slides →
PP — 02
Progress Presentation I
Presented: February 2025
View Slides →
PP — 03
Progress Presentation II
Presented: March 2026
View Slides →
PP — 04
Final Presentation
Scheduled: May 2026
Not Yet Available
Team

About Us

The researcher and supervisors behind LegalEase LK.

Researcher
VS
Vimansa D. A. S
IT21346668
BSc (Hons) in Information Technology
Specializing in Information Technology
Faculty of Computing, SLIIT
it21346668@my.sliit.lk
Supervisors
Supervisor
Ms. Uthpala Samarakoon
Senior Lecturer · Faculty of Computing, SLIIT
uthpala.s@sliit.lk
External
Samadhi Bandara
L.L.B. (Hons) Peradeniya  ·  Attorney-at-Law of the Supreme Court  ·  Public Notary  ·  Commissioner for Oaths  ·  Company Secretary
Contact

Contact Us

Get in touch with the project team.

Email
it21346668@my.sliit.lk
Location
Faculty of Computing, SLIIT, Malabe
Department
Information Technology, SLIIT
Send a Message