An innovative Flutter application that offers a sophisticated AI assistant interface, seamlessly integrating multiple language models such as Gemma, Llama, and Mistral. This highly personalized assistant is meticulously trained on my data, delivering accurate and insightful responses to queries about my work and profile.
- Flutter: Frontend development.
- ChromaDB: Vector database.
- FastAPI: API endpoint creation.
- Railway.com: Backend deployment.
- Docker: Containerization for deployment.
- GitHub Pages: Frontend hosting.
- Groq: Optimized inference for large language models.
- Cross-platform compatibility (Android, Windows, Web).
- Chat interface with three AI models.
- File handling capabilities.
- Google Sheets integration for feedback collection.
- Markdown rendering support.
- Max API calls limit.
To run this project, you will need to configure the following environment variables in a .env file. Refer to .env.example for more details:
GEMMA_API_KEYLLAMA_API_KEYMISTRAL_API_KEYFEEDBACK_WEB_URLSHEET_ID
Clone the repository:
git clone https://github.com/AnishaShende/Private-LLMNavigate to the project directory:
cd private_llmCreate a .env file based on .env.example with your API keys:
GEMMA_API_KEY=your_gemma_api_key
LLAMA_API_KEY=your_llama_api_key
MISTRAL_API_KEY=your_mistral_api_key
FEEDBACK_WEB_URL=your_feedback_url
SHEET_ID=your_google_sheet_idInstall dependencies and run the application:
flutter pub get
flutter runDesktop (Windows/macOS/Linux):
flutter build windows
flutter build macos
flutter build linuxMobile:
flutter build ios
flutter build apkWeb:
flutter build webDocker image for the API endpoint is available at: Docker Hub
flutter_markdown: ^0.7.5 - Markdown rendering.sidebarx: ^0.17.1 - Sidebar navigation.typewritertext: ^3.0.9 - Typewriter text effects.flutter_dotenv: ^5.2.1 - Environment variable management.groq_sdk: ^1.0.2 - AI model integration.shared_preferences: ^2.3.4 - Local storage.gsheets: ^0.5.0 - Google Sheets integration.http: ^1.2.2 - HTTP requests.url_launcher: ^6.2.5 - URL launching.window_manager: ^0.3.8 - Window management.open_file: ^3.5.10 - File handling.
Install the application using the installers available in the installers folder. For desktop platforms, follow the setup wizard to complete the installation.
The website has been successfully deployed and is accessible at this link .
For feedback, feel free to reach out to me at: anishaashende@gmail.com











