Agentic filesystem

[aniketmaurya]

Often a times we need to search in PDFs, if they are stored on a cloud like S3 then we have to download all the files and do search. This is wasteful.

An ideal way could be to have an Agentic filesystem with READ, WRITE, DELETE, SEARCH methods.

• Each write will also store a metadata with overview about the document.
• Read will read and cache the file to the disk
• Search will first look for the metadata (semantic search) and list all the possible document path and read them for the final text search.