Overview

Jawab is a specialized full-stack AI assistant designed to bridge Arabic and French linguistics. It provides users with expert-level translation, deep grammatical analysis (Iʿrāb), advanced conjugation, and text vocalization (Tashkil). Moving beyond standard chatbot templates, Jawab offers a highly customized, multi-modal user experience across both web and native mobile platforms.

Key Features & Achievements

• Specialized Linguistic AI: Engineered a stateful, agentic AI backend using LangGraph to handle complex, continuous conversations specific to Arabic and French linguistic nuances.
• Multi-Modal Capabilities: Designed the system to natively process diverse inputs. Users can interact via standard text, record voice audio for automatic transcription, or upload images for AI-driven OCR and context extraction.
• Bespoke Cross-Platform Frontend: Developed a fully custom, modern web interface using React and Tailwind CSS. By leveraging the Chainlit React Client library, the frontend breaks free from default chatbot UIs, providing a highly tailored and branded user experience.
• Native Android Application: Packaged the React frontend into a native Android mobile application using Capacitor, ensuring a seamless, app-store-ready experience with native device capabilities.
• Robust AI Infrastructure: Architected a production-ready backend with FastAPI and PostgreSQL. It incorporates intelligent routing between top-tier LLMs, strict observability and tracing through Langfuse, per-user API budget enforcement using LiteLLM, and object storage via MinIO.
• Secure Production Deployment: Designed and managed a secure production environment, configuring NGINX as a reverse proxy to handle isolated container networking, SSL termination, and secure routing across the full stack.