Forget clunky AI pins! Your smartphone or smartwatch already has the potential to be your personal AI powerhouse. This open-source project harnesses the full might of generative AI, multi-modal capabilities, and RAG (Retrieval Augmented Generation) to turn your existing devices into intelligent companions.
- Multi-Platform: Experience seamless AI assistance on iOS and Android, thanks to the versatility of Flutter.
- Voice Mastery: Communicate effortlessly with native Android STT/TTS or unlock even more languages with Google Chirp's advanced speech recognition.
- Personalized to You: Your AI remembers your preferences and past conversations, providing tailored recommendations and assistance through on-device RAG technology.
- Versatile Modes:
- Natural Voice Chat: Engage in dynamic conversations like you would with a friend.
- Translation Mode: Break down language barriers with real-time translation powered by Chirp.
- Multi-modal Exploration: Ask questions about anything you see through your camera, unlocking a new dimension of interaction.
- Leverage the Gemini 1.5 Flash & Pro Models: Experience cutting-edge AI language understanding and generation.
- Tap into Powerful Tools:
- Location-aware responses
- Real-time weather forecasts
- Sunrise and sunset information
- Web research through Tavily (API key required)
- Up-to-date fiat and crypto currency exchange rates
- Web search via DuckDuckGo Assist
- Business insights with Alpha Vantage (API key required)
- Song lyric lookup
- Expand your AI's toolkit with SerpAPI integration
- Calendar integration
- Email integration
- SMS / Text message, call, and Contacts integration
Transform your mobile device into a personalized AI assistant. Contribute to this open-source project and shape the future of AI on your terms!
Remember: API keys for Tavily and Alpha Vantage are required to access their respective features.
Let's build a smarter, more connected future together!
Notes:
- A demo video of the app: https://www.youtube.com/watch?v=kCtHH6XG5as
- This project was submitted to the Gemini API Developer Competition.
- Please find the related cloud functions - which comprise the backend - in the Open Multi-Modal Personal Assistant organization. Once the judging is over this repository will eventually be renamed and transferred there as well.
- The repository state at the time of the submission was the last commit on August 12th before I recorded the demo: https://github.com/CsabaConsulting/InspectorGadgetApp/tree/b7de11fa3d9122fd96b85e7c708802b371bb15c9
- After the Made By Google '24
someone may compare Open MMPA to Gemini Live, however we must notice that:
- Open MMPA targets embedded form factors
- Open MMPA also features a local Vector Database for history and personal RAG
- Open MMPA may lack some multi modal capabilities and integrations, but those are all planned
- Open MMPA is open source
Initial application source code scaffold was generated by the Very Good CLI 🤖