Question 1

How does PrivyAI guarantee 100% device-local privacy?

Accepted Answer

PrivyAI memory-maps pre-quantized GGUF neural weights directly inside your device's isolated application sandbox. All model token generation and neural calculations are processed purely on your physical hardware (Apple Metal or Snapdragon NPU). No outbound TLS network sockets or cloud telemetry are ever initiated during inference.

Question 2

How does the local OpenAI API proxy work with my laptop & scripts?

Accepted Answer

When enabled, PrivyAI spins up an embedded HTTP server directly on your handset exposing standard /v1/chat/completions and /v1/models endpoints. Any Python script, LangChain agent, Node.js app, or local IDE tool on your Wi-Fi or USB hotspot can connect simply by setting base_url='http://<your-phone-ip>:8080/v1'.

Question 3

Will running local LLMs drain my battery or physical RAM?

Accepted Answer

PrivyAI uses heavily optimized Q4_K_M and 1-Bit quantization to minimize RAM footprint and thermal output. The app includes built-in unified RAM diagnostics that measure available physical memory before loading weights, ensuring zero system instability or background throttling.

Question 4

What platforms and devices are supported?

Accepted Answer

PrivyAI supports iOS (iPhone 12 & newer, iPad Pro/Air with A14/M1 series or newer) and Android devices with 6GB+ RAM and Snapdragon 8 series / Tensor processors.

Question 5

Is PrivyAI compliant with HIPAA, GDPR, and legal privilege standards?

Accepted Answer

Yes. Because zero prompt data leaves your handset memory, PrivyAI eliminates data transit risks entirely. Regulated clinical teams, law firms, and financial institutions can process confidential records without requiring third-party cloud BAA signoffs.

PrivyAI Privacy Policy

1. Zero Cloud Data Collection

2. Local Application Sandbox

3. Regulatory Compliance

4. Contact Our Privacy Office