Web App Featured Project

Keyli - Court Transcription System

A revolutionary speech-to-text platform for the Maldivian judiciary, automating court transcriptions with custom AI models for the Dhivehi language.

Client
Department of Judicial Administration
Year
2023
Category
Web App
Keyli - Court Transcription System

Technologies Used

React JavaScript Python AI/ML Speech Recognition CourtConnect SSO

Project Tags

Government Judiciary AI Speech-to-Text React Python

Keyli is a groundbreaking speech-to-text platform developed for the Department of Judicial Administration (DJA) of the Maldives, designed to revolutionize the court transcription process. By leveraging custom-trained artificial intelligence models, Keyli accurately transcribes court proceedings from Dhivehi speech into text, drastically reducing manual effort and delays.

Project Overview

The Maldivian judiciary faced significant backlogs in producing court transcripts, a process that traditionally took weeks of manual work. To address this, the DJA partnered with Intellea to create Keyli, a digital solution that automates transcription. The system is built to understand the unique nuances and dialects of the Dhivehi language, marking a significant leap in the nation’s judicial digital transformation.

Core Features

AI-Powered Transcription

  • Custom speech-to-text models trained specifically for the Dhivehi language.
  • Handles various local dialects and legal terminology.
  • Reduces transcription time for a one-hour hearing to approximately 10 minutes.

Editor & Workflow

  • A web-based editor allowing court staff to review and correct the AI-generated transcript.
  • Playback of audio synchronized with the transcribed text.
  • Tools for formatting, speaker identification, and annotation.

System Integration

  • Secure single sign-on (SSO) integration with the judiciary’s ‘CourtConnect’ portal.
  • Role-based access control for judges, court administrators, and transcription staff.
  • Designed to fit seamlessly into existing judicial workflows.

Technical Architecture

Frontend

  • A modern, responsive web interface built with React and JavaScript.
  • Designed for ease of use by non-technical court staff.

Backend

  • Powered by a robust Python backend.
  • Serves the AI models and handles the audio processing pipeline.
  • RESTful APIs for communication between the frontend and backend services.

AI & Machine Learning

  • Core of the system is a set of custom-trained speech recognition models.
  • Models are continuously improved to enhance accuracy.

Impact and Significance

Efficiency and Speed

  • Reduced transcription time from weeks to mere minutes, clearing case backlogs.
  • Freed up court staff from laborious manual transcription to focus on higher-value tasks.

Modernizing the Judiciary

  • A key pillar in the digital transformation of the Maldivian judicial system.
  • Enhanced transparency and accessibility of court records.
  • Praised by the Chief Justice and the Speaker of Parliament as a revolutionary step.

Technological Advancement

  • One of the first successful large-scale applications of AI and speech recognition for the Dhivehi language.
  • Showcased local capability in developing complex, AI-driven solutions for public sector challenges.

Conclusion

Keyli stands as a landmark project in Maldivian legal technology. It not only solves a critical operational bottleneck for the judiciary but also represents a significant step forward in applying artificial intelligence to the Dhivehi language. The system enhances the efficiency and accessibility of justice in the Maldives, demonstrating Intellea’s expertise in creating high-impact, specialized solutions for government institutions.

← Back to Portfolio