Voice Freight Broker

Voice Freight Broker is an automated logistics management system designed to streamline the communication between freight brokers and truck drivers. It leverages AI-driven voice processing to handle routine check-ins and journey tracking, allowing brokers to manage shipments more efficiently.

The application serves as a comprehensive backend service and management dashboard for the freight industry. It features Automated Driver Check-ins, where the system handles voice calls with drivers to verify their location, load information, and status. It uses AI to analyze these conversations and automatically flag potential issues, enabling brokers to proactively address problems before they impact delivery schedules.

The system includes Journey Tracking functionality, managing the entire lifecycle of a delivery journey from origin through transit to final destination. It tracks stops and estimated arrival times (ETAs) in real-time, providing brokers with up-to-date information about shipment progress. The platform features Real-time Call Processing, integrated with voice AI, where the system processes live calls, generates transcripts, and stores recordings for broker review.

An Operational Dashboard provides a centralized web interface that allows freight managers to monitor all active journeys, view stop details (like nearest highways or delay info), and receive instant notifications regarding transit alerts. The system maintains Persistent Logistics Data, keeping detailed records of all interactions, journey states, and driver check-ins to ensure a complete audit trail for every shipment.

The project is built using a sophisticated AI orchestration and web development stack. The core intelligence is powered by LangChain and LangGraph, which manage complex AI workflows and state-based decision-making during driver interactions. The system uses Retell AI for real-time voice call processing and natural language conversation between the system and drivers. The backend is implemented with FastAPI, providing a high-performance, asynchronous web server to handle API requests and webhooks.

The system utilizes SQLAlchemy ORM for structured data management, with a SQLite database for persistent storage of journeys, stops, and call metadata. The entire environment is containerized using Docker and Docker Compose, ensuring consistent deployment across different platforms (Windows, Linux, and macOS). The frontend uses Jinja2 engine to render dynamic HTML templates for the transit dashboard and check-in interfaces. The system integrates Ngrok for webhook management, allowing the local development environment to securely receive real-time call events from external AI services. It employs sentence-transformers for text analysis and rouge-score for evaluating text similarity and call transcript accuracy.

• Developed automated logistics management system for freight brokers and truck drivers
• Implemented automated driver check-ins using AI-powered voice calls for location and status verification
• Built journey tracking system managing complete lifecycle from origin to final destination
• Created real-time call processing system with AI integration for live call handling and transcript generation
• Designed operational dashboard for freight managers to monitor active journeys and transit alerts
• Implemented persistent logistics data system maintaining complete audit trail for shipments
• Utilized LangChain and LangGraph for complex AI workflows and state-based decision-making
• Integrated Retell AI for real-time voice call processing and natural language conversations
• Developed FastAPI backend for high-performance, asynchronous API requests and webhooks
• Implemented SQLAlchemy ORM with SQLite database for structured data management
• Containerized application using Docker and Docker Compose for consistent deployment
• Utilized Jinja2 templating engine for dynamic HTML templates in dashboard interfaces
• Integrated Ngrok for secure webhook management and real-time call event handling
• Employed sentence-transformers and rouge-score for text analysis and transcript accuracy evaluation
• Worked in team environment following Agile methodologies for iterative development

Automated Driver Check-ins

Journey Tracking

Real-time Call Processing

Operational Dashboard