Lakra Documentation
Welcome to Lakra
Lakra is a comprehensive annotation and evaluation system for machine translation quality assessment, designed specifically for the WiMarka project. Built with React, TypeScript, TailwindCSS, and Supabase.
Lakra combines AI-powered assessment with human-in-the-loop evaluation to provide comprehensive quality analysis for machine translation, supporting multiple annotation workflows and quality scoring mechanisms.
Quick Links
Getting Started with Lakra - Start here if you’re new to Lakra
Installation Guide - Installation and setup instructions
Architecture - System architecture overview
API Reference - Complete API documentation
Key Features
Interactive Annotation: Text highlighting, error classification, and quality scoring
AI-Powered Evaluation: Automated quality assessment with confidence levels
Multi-Role System: Administrators, Annotators, and Evaluators
Voice Recording: Audio annotation support for corrections
Real-time Processing: Immediate feedback and validation
Comprehensive Analytics: Progress tracking and quality metrics
Table of Contents
User Manual