Hardware Key Encryption
Expo SecureStore native hardware encryption
Zero Middleware Logging
Direct client-to-LLM provider API requests
Complete Device Isolation
Chat histories stored 100% locally on device
Overview & Scope
Chatty AI is a privacy first multi model AI mobile application developed and maintained by Pramuda Kulathunga. Chatty AI enables users to securely connect their own AI service provider credentials (Google Gemini, OpenAI, Anthropic, DeepSeek, Groq, or custom LLM endpoints) to interact with cutting edge artificial intelligence without relying on centralized data scraping servers.
Information We Collect & Store
Chatty AI is engineered with a strict client side data model. The application stores only the minimum data required for prompt execution and UI state management directly on your physical device:
Encrypted API Credentials
API keys entered by the user are stored in the device's encrypted hardware keychain (`Expo SecureStore`) with fallback to local app storage. Keys are never transmitted to any third party other than the designated provider during API requests.
Chat Conversations & Logs
Message history, prompt text, AI responses, dynamic titles, and timestamps are stored locally on your device (`AsyncStorage`) and cached in high speed RAM for instant rendering.
Custom Provider Configurations
Custom OpenAI compatible base URLs, custom model identifiers, headers, and active provider selections configured within settings.
App Preferences & Themes
Selected visual theme palette, dark/light mode preference, dynamic voice settings, and custom response length preferences.
Purpose of Processing
- Executing user chat queries directly with configured AI model endpoints over secure encrypted HTTPS connections.
- Providing dynamic fallback routing to operational models when a model identifier (e.g. legacy Llama models) is deprecated or temporarily unavailable.
- Caching chat sessions in RAM and local storage to enable instant drawer navigation and offline history viewing.
- Formatting text-to-speech output and rendering interactive markdown code blocks for enhanced developer experience.
Security & Storage Architecture
All network transmissions between Chatty AI and official LLM provider APIs enforce 256-bit TLS/SSL encryption. Sensitive API keys are hardware encrypted on native mobile devices via Expo SecureStore. Because Chatty AI operates without an intermediary backend server, your conversations and API credentials never pass through any developer owned tracking server.
Supported AI Providers & Privacy Policies
When sending prompts, your request is transmitted directly to the API endpoint of your selected provider. You can review each AI provider's privacy and data retention policies here:
User Control & Data Erasure
You maintain complete, immediate control over all data. You can wipe all chat histories, remove individual chat sessions, and clear saved API keys directly from the app drawer and settings screen at any time. Because no user data is stored on remote servers, clearing app data or uninstalling the app permanently purges all stored information. For any questions or support, contact the developer at pramudakulathunga@gmail.com.
Developer Contact Details
Pramuda Kulathunga
Computer Engineering Undergraduate, University of Jaffna
pramudakulathunga@gmail.com