TechNest Logo

TechNest Service

PrivyAi

← All Policies
PrivyAi Logo
Official Policy

PrivyAi Privacy Policy

Last Updated: July 18, 2026

This Privacy Policy explains how TechNest Service Ltd (“we”, “us”, or “our”) manages, processes, and protects your information when you use our mobile application, PrivyAi (the “App”).

We are committed to maintaining the highest standards of user privacy. PrivyAi is built as a local-first, privacy-by-design application. By design, the App processes and retains all conversations, prompts, and personal data locally on your physical device. We do not collect, transmit, share, or store your private data on remote servers.

Please read this Privacy Policy carefully to understand our practices regarding your data and how we satisfy the compliance standards of the Apple App Store and Google Play Store.

1. Zero Data Collection Policy (Zero-Cloud)

We believe that your personal conversations and data belong solely to you.

  • No Chat Logging or Transmission: We do not log, review, monitor, or upload your prompts, messages, questions, or responses. All inference (processing) is done locally on your device hardware (CPU/GPU/RAM).
  • No Personal Accounts: You do not need to register, create an account, verify an email address, or provide any personal information to use the App.
  • No Telemetry or User Analytics: We do not track your location, behavior patterns, search terms, or device usage statistics. There are no tracking scripts, cookies, or analytics SDKs embedded in the App.

2. On-Device Storage and Sandboxing

All data generated during your sessions resides exclusively inside your device’s secured application sandbox:

  • Local SQLite Database: Your chat histories, prompt inputs, and conversation logs are stored in a local SQLite database file inside the App’s secure sandboxed directory. This database is encrypted by default if device-level encryption (iOS Data Protection / Android File-Based Encryption) is enabled.
  • Local Model Inference: The App loads Large Language Model (LLM) files (GGUF format) into the device’s volatile memory (RAM) and runs math computations locally. Your queries are never sent to external AI servers for completion.
  • Context Compression & Summarization: To optimize device performance and memory usage, long chat histories are summarized locally. The raw historical messages are permanently discarded from the local database, and only the summary is kept on-device.

3. Network Access and External Services

Although PrivyAi is designed to function entirely offline, certain optional features require secure network requests:

  • Model Downloads: To load models (e.g., Qwen, Gemma, Llama, Bonsai), the App initiates connections to public hosting services (such as Hugging Face). These requests are governed by the hosting platform’s privacy policy. The App only shares standard HTTP download request headers (such as IP address and User-Agent) necessary to retrieve the files.
  • Local API Server Proxy: The App contains an optional feature that starts a local HTTP API server (Port 8080) on your private network, enabling other local devices to access the LLM.
    • This feature is disabled by default.
    • It requires an active authorization token.
    • It operates strictly on your local Wi-Fi/subnet network. No traffic is routed over the public internet, and no external incoming connections are accepted.
  • Firebase Services: We use Firebase Core and Firebase Remote Config solely to download update catalogs of supported model metadata (file links, model sizes, and RAM specifications). No personal identifiers, IP logs, or chat data are associated with or transmitted to Firebase.

4. Platform Permissions

To deliver on-device AI functionality, the App requests specific standard permissions from your operating system:

  • Notifications (iOS / Android): Requested to show background model download progress, download completions, and task scheduler updates.
  • Local Network / Multicast (iOS / Android): Requested only if you enable the Local API Server Proxy feature. This is required to allow the local server to bind to your private Wi-Fi network interface and accept local subnet connections.

5. Data Retention, Control, and Deletion

Because we do not collect or store your data on our servers, we have no means of retaining or deleting it on your behalf. You have total control over your local data:

  • Clear Chats: You can delete individual messages or clear entire chat sessions instantly within the App’s interface. Once deleted, they are permanently expunged from the local SQLite database.
  • App Uninstallation: Uninstalling the App from your device permanently deletes the application sandbox, including all local databases, settings, API configurations, and downloaded LLM model files. This action is irreversible.

6. Global Legal & Store Compliance

We ensure full compliance with international data privacy frameworks:

  • COPPA (Children’s Online Privacy Protection Act): The App does not collect, solicit, or maintain personal information from individuals under the age of 13.
  • GDPR (General Data Protection Regulation): Because we do not collect, process, or transmit personal data of European Union residents, our data collection footprint is zero. Your data rights (Access, Portability, Erasure) are fully exercised by you directly through local data deletion or app uninstallation.
  • CCPA (California Consumer Privacy Act): We do not sell, trade, or share any personal information. We do not collect personal identifiers from California consumers.

7. Changes to This Privacy Policy

We may update this Privacy Policy periodically. We will notify you of any modifications by updating the “Last Updated” date at the top of this policy and publishing the update inside the App. We recommend reviewing this policy regularly to stay informed of our privacy practices.

8. Contact Us

For questions regarding this Privacy Policy or our local-first practices, please contact us: