OpenTeller Protocol (v1.0) 🏦
A Privacy-Preserving, Voice-First Banking Interface for the Visually Impaired. > Compliant with RBI Circular 2015 & IBA Standards 2013.
Live Link: (https://openteller-protocol.streamlit.app/)
🚨 The Problem
For the 2.2 billion visually impaired people globally, using an ATM is a security nightmare.
Shoulder Surfing: They cannot see if someone is peeking at their PIN.
Hardware Reliance: They struggle to find the specific keys on non-standardised machines.
Privacy Loss: They often have to ask strangers or guards for help, compromising financial autonomy.
⚡ The Solution: OpenTeller Protocol
OpenTeller is a software protocol that transforms any standard banking terminal into an Autonomous Assistive Interface. It uses Edge AI (Computer Vision & Voice) to replace physical reliance with conversational trust.
Key Capabilities
👁️ Sentinel Vision (Shoulder Surfing Detection): Uses YOLOv8 to monitor the camera feed. If a second person enters the frame behind the user, the transaction pauses immediately.
🛡️ Privacy Vault (DPDP Act Compliant): A PII-Redaction layer that scrubs sensitive data (Aadhaar/PAN) before it touches any AI processing unit.
🎧 Hardware Handshake: Detects the insertion of a 3.5mm Headphone Jack to toggle between "Public Mode" (Visual) and "Privacy Mode" (Voice-Only), as mandated by IBA Standards.
🏛️ Compliance Architecture
This protocol was architected to strictly adhere to Indian Banking Regulations.
Regulatory Body
Mandate/Circular
OpenTeller Implementation
RBI (Reserve Bank of India)
Master Circular DBR.No.Leg.BC. 21/09.07.006
Talking ATM Mandate: 1/3rd of ATMs must be accessible. OpenTeller provides the software layer for this.
IBA (Indian Banks' Assoc.)
Standards on Accessible ATMs (Feb 2013)
Hardware Trigger: Logic implemented to blank the screen and activate voice only when the headset is detected.
NBC (National Building Code)
Part 3, Annexe B (Accessibility)
High Contrast UI: Interface uses specific Black/Yellow contrast ratios for Low Vision users.
🛠️ System Architecture
The system runs purely on Edge Compute to ensure <200ms latency and data privacy.
graph TD
User((👤 Visually
Impaired User)) -->|Voice Command| UI[🎤 Interface Layer]
User -->|Presence| Cam[📷 Camera Feed]
subgraph Edge_Device [Local Processing Unit]
UI -->|Audio Stream| PII[🛡️ Privacy Vault]
Cam -->|Video Frames| YOLO[👁️ Sentinel Vision]
YOLO -- Intruder Detected -->|STOP Signal| UI
PII -- Sanitized Text -->|JSON Intent| Core[🧠 Banking Core]
end
Core -->|Transaction Success| TTS[🔊 Text-to-Speech]
TTS -->|Audio Feedback| UI
🚀 Installation & Usage
Prerequisites
Python 3.8+
Webcam (for Sentinel Vision)
- Clone the Repository
git clone https://github.com/YOUR_USERNAME/OpenTeller-Protocol.git cd OpenTeller-Protocol
- Install Dependencies
pip install -r requirements.txt
- Launch the Protocol
streamlit run src/interface/app.py
🌐 How to Deploy (Live Demo)
This project is optimized for deployment on Streamlit Community Cloud.
Push this code to a public GitHub repository.
Go to share.streamlit.io.
Connect your GitHub account.
Select the OpenTeller-Protocol repository.
Set the Main file path to: src/interface/app.py
Click Deploy.
The link generated (e.g., https://openteller-protocol.streamlit.app/) serves as the live portfolio demo.
📅 Roadmap (Version History)
v0.1 (MVP) - Initial Release
[x] Core Voice-to-Text Engine.
[x] Basic PII Redaction Logic.
[x] High-Contrast UI (Yellow/Black).
v1.0 (Current) - Compliance Patch
[x] IBA Hardware Simulation: Added logic to detect Headphone Jack insertion.
[x] Sentinel Vision: Integrated YOLOv8 for real-time intruder detection.
[x] Accessibility Fix: CSS overrides to ensure black-on-yellow text compliance.
👨💻 Attribution
Architected by: Sumanta Pani Product Manager & Lead Developer
Built to bridge the gap between AI Innovation and Financial Inclusion.