Similar Notes

by Young Lee
favorite
share
0.0
(0)
5
4
3
2
1
Score: 45/100
Description

The Similar Notes plugin finds and recommends notes that are semantically related to what you're currently writing, using language models for real-time analysis. It displays up to five similar notes at the bottom of your note, making it easier to spot hidden connections in your ideas without relying on exact keyword matches. The plugin is self-contained with its own vector database and gives you the choice between built in Hugging Face models or connecting to Ollama for custom embeddings. No external cloud API is needed, and it smartly ignores notes already linked from your current document.

Reviews
No reviews yet.
Stats
61
stars
4,496
downloads
5
forks
148
days
4
days
4
days
1
total PRs
0
open PRs
0
closed PRs
1
merged PRs
27
total issues
10
open issues
17
closed issues
0
commits
RequirementsExperimental
  • Internet connection for initial model download

  • Local Ollama installation for custom embedding models (optional)

Latest Version
4 days ago
Changelog

Fixed

  • Skip Non-Markdown Files: Binary files (images, PDFs, etc.) no longer trigger unnecessary embedding generation
    • Previously, opening an image or PDF would attempt to embed the binary data as text
    • Now only .md files are processed for similar notes lookup
    • Eliminates wasted API calls and meaningless search results

Improved

  • Ollama Bug Workarounds: Added automatic detection and workarounds for Ollama v0.12.5+ bugs
    • Auto-detect max token limits to avoid random embedding failures
    • Conservative token counting (3.5 chars/token) for better payload estimation
    • Sequential processing to match Ollama's server-side queuing behavior
  • Duplicate Embedding Prevention: Check repository before generating new embeddings
    • Avoids regenerating embeddings for already-indexed chunks
    • Reduces unnecessary API calls during similar notes lookup
  • Enhanced Logging: Comprehensive logging for troubleshooting Ollama issues
    • Payload sizes, token counts, and timing information logged
    • Helps diagnose embedding failures and performance issues
README file from