It does this by generating embeddings using a local model (supabase's gte-small model) and finds similar posts with @OramaSearch, a local vector DB.
Show HN: Nomad Hypertext, a local first writing app built around semantic search
github.com
1–8 of 8 posts
It does this by generating embeddings using a local model (supabase's gte-small model) and finds similar posts with @OramaSearch, a local vector DB.
Show HN: Nomad Hypertext, a local first writing app built around semantic search
github.com
I really love this app/concept. I'm curious from your time working on the app what applications/use cases beyond note-taking you might have uncovered, that might not be obvious at first glance? (e.g. If studying a topic, creating a note of questions to quiz yourself from time to time, with the answers likely being found in semantic search.)
I generally don't take notes when trying to learn something, I usually like to be more hands on (by doing problems, making a project applying the principles, etc.) When I write it's usually to reflect, or to hash out a more open ended problem.
This looks hella cool. Let me ask you - could you use this (or the sister project I see, Yurt) to upload a novel, or some large block of text and follow that around. Eg... take Moby Dick from gutenberg and follow each paragraph around?
I tried doing this with Yurt to deploy the first chapter of Understanding Media by Marshall McLuhan: https://themediumisthemessage.nicholaschen.io/posts/about.tx...
The main challenges are chunking the text so that the paragraphs actually encapsulate a concrete idea, so that the connections between them are valuable.
I really like this concept, echoing what others have said. I'm an avid user of Pocket and tagging is wonderful, but auto-tagging and periodic surfacing is something I've wanted forever. Doing it at the note/db level is awesome