Hello DEV Community! 👋
I'm excited to share SupplierSignal AI, a project developed by our team Code Killers for Hacktoberfest.
👨💻 Team Members
Sayyad Abu Basha — Team Lead
Veeru Naidu
NVS Manikanta
💡 The Problem
Small online sellers often manage supplier commitments through messages while invoices and received-stock records come in different formats. Manually comparing these records takes time, and discrepancies in product details, prices, or quantities can go unnoticed, potentially leading to financial losses and supplier disputes.
🛠️ Our Solution
SupplierSignal AI is a mobile-friendly, local-first web application designed to help small sellers verify supplier commitments against invoices and received goods.
The application aims to:
🤖 Use Gemma, an open-weight AI model, to extract structured information from purchase records.
📋 Compare supplier commitments, invoice details, and received-stock information.
🔍 Flag potential discrepancies, missing information, and uncertain matches.
📄 Generate discrepancy reports to support further investigation.
🔐 Keep sensitive business records under the seller's control through a local-first approach.
The goal is to make purchase verification more organised, transparent, and efficient.
⚙️ How It Works
Create a purchase case.
Add the supplier agreement or message.
Enter the invoice details.
Add the goods actually received.
Analyse the records and review the resulting report.
🧰 Technology
Python — backend logic
Gemma — open-weight AI-assisted information extraction
Ollama — local model runtime
Web technologies — user interface and application workflow
🎥 Demo and Source Code
🔗 GitHub Repository: https://github.com/abu-124/SupplierSignal-AI
📹 Demo Video: [Add your demo video link here]
🌱 What We Learned
Building SupplierSignal AI helped us explore how open-weight AI can be combined with application logic to address a practical business problem. We also learned about structuring records, comparing extracted information, and presenting potential discrepancies in a useful report.
This is an evolving prototype, and we aim to improve extraction accuracy, evidence traceability, and the review workflow in future iterations.
🙌 Feedback Welcome!
We would love to hear your suggestions. What additional checks would be useful for small sellers—for example, product-size mismatches, duplicate invoices, or delivery-date verification?
Thank you to the DEV and MLH communities for creating opportunities to learn, build, and share open-source projects!
Hacktoberfest #DEVChallenges #AI #Gemma #Python #OpenSource[](
)
Top comments (0)