Core concepts
These terms appear throughout Curator. Each has a dedicated page below.
| Term | What it means |
|---|---|
| Storage Source | Where your files physically live. A folder on this Mac or a cloud bucket you own. Curator reads the files in place and never moves or alters them unless you ask. |
| Ollama | A free, open-source app that runs AI models directly on your Mac. Curator uses it so everything stays local. |
| Model | A single AI model that runs inside Ollama. You download (pull) the ones Curator needs. |
| Processing | The work that makes a file searchable. Curator extracts its text, writes a description, and prepares it for search and chat. |
| Embeddings | A numeric representation of meaning. This is what lets Curator match ideas, not just exact words. |
| Semantic search | Search by meaning instead of by exact filename or text. |
| Clusters | AI-built smart folders on the Organize screen. They are virtual, so no files are moved. |
| Profile | A local, name-only identity. No account, no email, no cloud. |
| Jobs | The background task queue. Processing, organizing, and syncing all run here. |
Dig deeper
Section titled “Dig deeper” How Curator works The whole app in three plain ideas.
Storage sources Where your files live and how Curator reads them.
Ollama and models The local AI engine and the models it runs.
How processing works What happens to a file to make it searchable.
Profiles Local, name-only identities.