stowboard

What it does

Stowboard is your ultimate personal assistant when it comes to storing photos or bookmarking websites or making random notes at the spur of a moment. As people we tend to capture photos or bookmark sites or note down an address quickly in a notes application for later retrieval, but as time passes on we find it tough to retrieve it back. Stowboard solves this by automatically analysing the content you store and tags them so that you can perform searches like - get me the addresses that i saved or get me the photos of the cutlery i took or get me the wallet website i saved making it much for easier for the user to locate their content instead of endlessly sifting through multiple apps. When storing images the app makes use of Gemini's vision capability to identify objects within the image to create tags and summary. With links the app extracts the title, keywords and description and feeds it to the text generation capability of Gemini to create tags and summary. With text the app uses Gemini to categorise the text into address, mobile number, notes or reminders and creates a tag and summary from this. The tags and summary generated by Gemini is then converted to vector using Gemini's embedding capabilities and is stored in Redis. When the user searches for something the app converts the text into vector embedding using Gemini's embedding capabilities and does a vector based search. The app also uses Firebase auth, Firestore and Firebase cloud functions.