Udhaar-Khata: A Voice-Powered Local AI Ledger for Small Businesses

Developer Ishan has created 'Udhaar-Khata,' an open-source digital ledger designed to help small business owners manage customer credit without relying on cloud services. Built for his father, a shopkeeper in India, the application uses local open-weights models to solve common pain points like manual bookkeeping and privacy concerns. By integrating OpenAI’s Whisper for speech-to-text and Google’s Gemma 2 (2B) for semantic parsing, the tool allows users to record transactions in colloquial Hindi or Hinglish via voice commands. The system runs entirely offline on consumer hardware, ensuring data privacy and eliminating subscription costs. It features automated, culturally nuanced payment reminders and a simple interface for tracking debts. This project demonstrates the practical utility of running lightweight, open-source AI models locally to provide sovereign, cost-effective, and accessible solutions for non-technical users in real-world environments.
This is a summary. Read the full article at the original source:
Dev.toRelated stories
How we deploy AI agents to automatically fix production websites
The author describes the evolution of an AI agent designed to manage the MODX CMS. While the first part of the project focused on the agent understand…
How Cursor replaced my Stack Overflow, accelerated debugging, and brought back interest in pet projects
The author shares their personal experience of transitioning from traditional development methods to using AI tools, specifically the Cursor IDE. Init…
Building an Agent That Can't Afford to Be Wrong: Quran Sanity Agent
Developer Omar Afifi has launched the 'Quran Sanity Agent,' a specialized AI research tool designed to eliminate hallucinations in high-stakes religio…


