Last updated: May 1, 2025
H.M.L.D.L Senevirathna (“we,” “our,” or “us”) built the FraudShield mobile application as a free tool for detecting financial fraud using lightweight large language models (LLMs) deployed directly on users’ devices. This privacy policy describes how we collect, use, and protect your information when you use the FraudShield app.
FraudShield does not collect, transmit, or store any personal or sensitive user data on external servers. The entire fraud detection process—including text analysis—is performed locally on your mobile device.
The application:
Does not require account creation
Does not collect your device identifier, location, or contacts
Does not transmit messages, transaction data, or feedback to any server without consent
With your explicit consent, the app may collect anonymized feedback on classification results (e.g., false positives/negatives) to help improve model accuracy. This data is stored temporarily and securely for model updates.
When selecting or updating a language model within the app, the model file (in ONNX format) is downloaded from a trusted third-party repository (e.g., Hugging Face). No user data is sent during this process.
FraudShield is designed to work offline after model download. All analysis, detection, and fraud classification happen entirely on your device, ensuring compliance with data protection laws such as:
General Data Protection Regulation (GDPR) – EU
California Consumer Privacy Act (CCPA) – USA
We do not use cloud-based processing for detection or storage purposes.
To ensure your data remains private and secure:
We use ONNX quantized models optimized for mobile inference
We implement GDPR-compliant data retention policies (e.g., no background logging)
Any temporary data is stored in secure, app-isolated storage
You have full control over:
Whether to submit anonymized feedback (opt-in only)
Which fraud detection model to download and use
Deleting app data by uninstalling the app
FraudShield does not integrate with advertising SDKs, analytics tools, or third-party APIs that collect user data.
However, models are downloaded from trusted third-party hosting (e.g., Hugging Face). Only model files are retrieved—no personal data is transmitted.
FraudShield does not target or knowingly collect data from individuals under the age of 13. If you are a parent or guardian and believe your child has submitted data, please contact us to have it removed.
We may update this policy occasionally. Any changes will be posted here. Users are encouraged to review the Privacy Policy periodically for updates.
If you have questions or concerns regarding this privacy policy or your data privacy, please contact:
H.M.L.D.L. Senevirathna
Email: lakpriya@codescale.lk