M
Marian
AI
Marian runs models locally by default. Point it at your local model server.
Provider
Where inference and embeddings run.
Model endpoint
Base URL of the local model API.
Not detected — start Ollama at http://localhost:11434, then press Test.
Local models unavailable
Once Ollama is running at http://localhost:11434, your installed models appear here to switch and pull. Marian keeps working with cloud models meanwhile.
Auto-suggest links
Analyze notes on save and surface candidate links.
Extract entities
Detect people, places, projects & concepts.
Capture organize
Auto-files captures, or shows a before→after preview to confirm.
Tip: ⌘K searches notes or runs a command (type / for commands). Press ? for help.