I’ve always dreamed of importing my documents into Obsidian, having them automatically indexed, and then using a local LLM (copilot) to chat and conduct research. I found Omnisearch to be excellent because it extracts content from PDFs. However, this indexing is only accessible within Omnisearch. Currently, the only way to extract information from a PDF is through a manual action with the extraction plugin.
Do you have a solution to automate the creation of notes (text extracted from PDFs) for all PDFs/images in the vault at once? This would enable me to use Copilot with the content from the PDFs. Thank you very much.
i wouldn’t hold my breath on this one…
you’d need some serious scripting capabilities…you’d need to install some rigrep-all plus maybe a tesseract combo that talk to each other and a script that markdownifies the extracted texts
people read pdfs, annotate them and import highlights, notes into obsidian (see about zotero and zotero integration)
the first-hand experience of reading and musing over what you are reading cannot be replaced with ai “work”…