FileJIT
FileJIT is a file search and retrieval system that leverages AI-powered embeddings and similarity search to classify and organize files into a file directory, as well as find the most relevant files based on conversational prompting.
Features
File Embedding Generation
Converts file contents into dense vector embeddings using a pre-trained SentenceTransformer model.
Similarity Search
Uses FAISS to perform fast and efficient similarity searches for the closest file matches.
Top-N Results
Returns the top 3 most relevant files for a given query.
Dynamic File System Scanning
Automatically scans a specified directory for files and processes their contents.
Persistence
Supports saving and loading FAISS indices for reuse without regenerating embeddings.