← Back to all projects

Mnemosyne

in progress

Python Ollama

Mnemosyne screenshot

Named after the Greek goddess of memory, Mnemosyne (/niˈmɔziniː) is a RAG agent to complement my Obsidian knowledge base.

The application ingests my Obsidian markdown files, generate semantic embeddings and store them in a vector database. I can then submit a query to Mnemosyne to answer any questions that I have based on the knowledge available in my Obsidian vault.

At the moment, Mnemosyne runs on a local LLM which is Ollama. Embeddings are generated with nomic-embed-text.

View source on GitHub →