DukaanFlow AI: A Local-First AI Order Desk for Tiffin Sellers
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle . What I Built DukaanFlow AI is a local-first AI order desk designed for small tiffin services and home-food businesses. These businesses often receive customer requests through WhatsApp messages and phone calls. Managing daily changes manually can lead to mistakes, especially when a customer wants to skip one…
Title: DukaanFlow AI: AI Order Management Tool for Local Tiffin Services
This submission is for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle. The project, DukaanFlow AI, is an AI-powered order desk tailored for small tiffin services and home-food businesses. These businesses typically receive customer requests via WhatsApp messages and phone calls. Managing daily changes manually can result in errors, particularly when customers request exceptions like skipping a meal or canceling an order for a specific day.
DukaanFlow AI investigates how AI can assist sellers in interpreting such natural-language messages and converting them into structured order actions while maintaining seller control over important changes.
The primary audience for this tool includes independent tiffin providers, home-based food sellers, and small meal-subscription businesses. The solution aims to reduce the manual effort involved in managing orders, necessitating validation of pricing with real users. The live demo can be accessed at https://utkarshydv27.github.io/DukaanFlow-AI/.
Key components of the project include:
1. Local-first AI architecture using Ollama to run a language model locally.
2. Qwen2.5 1.5B for local natural-language interpretation.
3. Rule-based fallback handling for predictable behavior.
4. Seller confirmation before applying significant order changes.
5. GitHub Pages for deploying the public demo website.
6. The code is available in the GitHub repository at https://github.com/utkarshydv27/DukaanFlow-AI.
The project emphasizes a local-first approach, reducing reliance on paid, per-request cloud AI APIs. This approach also minimizes the need to transmit customer messages to external AI providers, ensuring that processing remains local. However, local inference necessitates appropriate hardware, electricity, and maintenance, with privacy depending on the complete implementation. The primary goal is to empower developers to customize AI for specific real-world problems, such as the unique needs of small food sellers.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
