About
Turn your notes into knowledge.
An idea jotted down, pages of notes, a subject to understand. Luma exists to turn those raw notes into clear, structured knowledge.
Stay in control.
A simple approach leaves room for what matters: understanding and learning. You keep your original notes and can read through, correct, or rewrite each chapter.
Luma offers a structure. You decide what it becomes.
A path to learning.
It all starts with your notes. They become editable chapters, still linked to the original passages. Then flashcards and questions help you review at your own pace.
Choose your AI.
Codex is the recommended engine for its speed. It requires an internet connection and sends your notes to the service associated with your account.
Ollama offers a local alternative: once it and its model are installed, it can process your notes on your Mac. Wikipedia search and GitHub backup remain optional.
Detailed technical requirements
With Codex
Codex CLI must be installed and signed in to your ChatGPT account. An internet connection is required to use it, and your account limits apply. Turning off Wikipedia search in Luma does not disable Codex’s connection to the remote service.
With Ollama
Ollama must be installed and running on your Mac. The llama3.2:3b-instruct-q8_0 model must be downloaded separately from Luma and takes up around 3.4 GB.
The initial setup and model download require an internet connection and disk space. Once everything is installed, local processing can work offline, without Wikipedia search or GitHub backup.