It works through an intuitive Telegram bot: the user simply takes a photo of any invoice or receipt and sends it. The system handles the entire process in the background, from extracting key information to organizing it in accessible databases, freeing the user from manual data entry and reducing errors.

The heart of the system is a serverless architecture that uses an AWS Lambda function connected to a webhook. When the bot receives an image, a GPT agent scans it to extract relevant data such as the amount, date, and expense concept, returning the information in a JSON format. The original image is archived in Google Drive, and the structured data is saved to Supabase and Google Sheets, offering a dual layer of storage and accessibility. Finally, the bot notifies the user of the operation's status.

Bill Tracker eliminates the need for manual invoice data entry, saving time and minimizing errors. By automatically centralizing and organizing expense information, it provides detailed financial control.

Bill Tracker's growth potential is enormous. A predictive analysis system could be integrated to generate personalized expense reports or detect anomalies. Additionally, functionality could be expanded to include automatic bank reconciliation or the ability to generate tax reports, turning the tool into a more complete financial solution. Integration with other messaging platforms like WhatsApp could also significantly increase its accessibility and utility.