Recallo is a full-stack AI application and browser extension that allows users to save web snippets, notes, and memories into a high-dimensional vector database. Using Temporal RAG (Retrieval-Augmented Generation), Recallo can answer questions based on your saved context while filtering by time and relevance to eliminate AI hallucinations.
- Temporal RAG Pipeline: Context-aware AI querying with time-filtered semantic search.
- Browser Extension: Instantly capture and embed knowledge directly from the web.
- Mind-Vault Dashboard: A sleek React interface to manage and interact with your digital memory.
- Local Embeddings: Utilizes
AllMiniLmL6V2EmbeddingModelrunning locally to convert text into vectors with zero API costs. - High-Speed AI Generation: Powered by Groq/Llama 3.1 and LangChain4j for lightning-fast inference.
- Backend: Java Spring Boot (v4.0.5), Maven, LangChain4j
- Frontend: React, Vite (Mind-Vault UI)
- Database: PostgreSQL, Supabase,
pgvectorextension - AI Models: Groq (Llama 3.1), Gemini API, Local MiniLM (Embeddings)
- Java 17+
- Node.js & npm
- A Supabase project with
pgvectorenabled
- Clone the repository:
git clone [https://github.com/Manas0110-01/recallo.git](https://github.com/Manas0110-01/recallo.git) cd recallo