Ollama Guide
How to use Ollama with a private notes app.
Prepare an Ollama model, configure it in Hibiscus, and test a note summary. Check the processing location, output, and AI edit history.
7 days. No account or card required. macOS 14 or later.
Direct answer
How do you use Ollama for notes?
Run Ollama with a model you choose, point Hibiscus at that runtime, select Ollama as the AI route, and run an approved transformation inside a note.
- Configure Ollama
- Select the route
- Run note actions
Prepare Ollama separately
Install Ollama, download a model that fits your Mac, and confirm that it responds. Hibiscus connects to the runtime you configure; installing Hibiscus does not automatically install an Ollama model.
Connect your model
Open AI provider settings, configure your Ollama endpoint, and select an available model. Use a runtime on the same Mac when you want request processing to stay local.
Try one clear request
Start with a short, non-sensitive note. Ask for a summary and next steps. Check whether the output preserves source facts before trying a longer document or a multi-step pipeline.
Review and restore
Check names, numbers, and conclusions. Open AI edit history to compare the earlier text and restore it when useful. A local model still needs review; its location does not guarantee accuracy.
If the connection fails
Confirm Ollama is running, the configured address is reachable, and the selected model is installed. If requests are slow, try a shorter note or a model that needs fewer resources.
Clear answers
Common questions.
Does Ollama send notes to cloud AI?
A normal local setup processes requests on the machine running your configured runtime.
Can I switch providers later?
Yes. Hibiscus does not force one provider.
Are AI edits reversible?
Yes. Hibiscus keeps AI edit history and supports restores.
