AI & intelligence

Where each AI action lives, Review Links, New Note from a Prompt, typing suggestions, Ask Library and providers.

On-device by default

The intelligence features use Apple Intelligence, running on your own device — Mac, iPhone or iPad. Your notes are processed locally and are not sent to us or to any third party. These features appear only on hardware where Apple Intelligence is available and enabled in System Settings.

Everything else in HelloNotes works without them — the AI is an addition, never a requirement.

Note actions

These live in the Note menu, beside Rename and Duplicate — not in a panel of their own. Each answer appears in the inspector tab that already holds that kind of information, so you find it where you would have gone looking anyway. On iPhone and iPad they are in the menu in the toolbar above the note.

  • Summarise Note — a short summary of the current note, shown at the top of the Outline tab. A summary is an outline at a different resolution.
  • Suggest Tags — proposes tags based on the content, in the Tags tab; you choose which to keep.
  • Suggest Links — finds notes that are plausibly related, in the References tab beside the backlinks they would join.
  • Rewrite or Expand Note… — improve, shorten, simplify, restructure or flesh out the note. You see the result and choose Replace Note or Insert Below; nothing changes until you do.

Selecting text gives you the actions that need your notes to make sense — link the phrase, find related notes, explain it using what you have written. On the Mac they appear in a small floating bar; on iPhone and iPad they join the standard selection menu. General rewriting is left to Writing Tools, which the system already puts in every text view.

Can’t remember where a command lives? On the Mac, ⇧⌘P opens acommand palette: type a command’s name to run it.

Note ▸ Review Links… (⇧⌘L) walks the note phrase by phrase the way a spell check does — Link, Skip or Never. You see the phrase in its own sentence and the opening lines of the note it would point at, because the same phrase can deserve a link in one paragraph and not the next.

  • Every phrase that names another note is a candidate. When a note turns up more than ten, the ten whose targets are most related to what you are writing are the ones shown — a long note can otherwise produce dozens, and a list nobody finishes reading is a list that gets skipped.
  • Never is remembered for that collection, so a phrase you have rejected stops coming back. It is kept on your device rather than in your notes — a decision you made is not something to commit to a shared repository.
  • Nothing is written until you finish, and the whole set undoes in one step.

This one needs no AI provider at all: it is an exact scan of your own text.

New note from a prompt

File ▸ New Note from a Prompt… (⌃⌘N) writes a new note, in two modes. Write composes one from a description. Research investigates a question on the web across several angles and lands the answer as a note, with its sources gathered at the end and a record of the question it came from.

Either way the draft is connected to what you already have: notes you own on the topic are offered as link targets, and every link that comes back is checked against your collection. Anything naming a note that does not exist is turned back into ordinary text, and the draft says so. A link that looks real but points nowhere is worse than no link.

You read the whole draft — title and body, both editable — before anything is created. Cancel leaves nothing behind. The button is Create Note.

Suggest as I type

Grey text appears after the cursor when you pause at the end of a line. ⌥⇥ or → accepts it; Esc dismisses it.

The suggestion is not part of your note until you accept it. It is drawn on screen and nothing more — it cannot reach a save, a search result, your links or a Git diff.

It runs on your device only: a suggestion that has to cross the network arrives after you have already typed past it. Off until you turn it on, in Assistant Settings ▸ Inline completion. Mac only for now.

Ask your library

Ask Library answers questions using your own notes as the source. It finds the relevant ones, reads them, and answers with citations back to the notes it drew on so you can check the reasoning and jump to the original.

It is most useful for questions you know you have written about but cannot place — “what did I decide about the pricing model?” rather than general knowledge.

The Assistant

The Assistant is a chat window with tools: it can search your collection, read notes, propose edits, and — if you enable them — search and fetch from the web.

Edits always require your approval. When the Assistant wants to change a note, it shows you the exact change first and waits. Nothing is written to disk until you approve it.

Skills let you save reusable instructions for repeated jobs, and deep research runs a longer multi-step investigation across your notes and, optionally, the web.

Using your own model

If you would rather not use Apple Intelligence — or your Mac does not support it — you can configure a different model in Settings ▸ AI.

Local, nothing leaves your machine:

  • Ollama — run ollama serve; no key needed.
  • LM Studio — start its local server; no key needed.
  • MLX — runs a downloaded model on Apple silicon.

Cloud, using your own API key:

Anthropic, OpenAI, Gemini, Mistral, Groq, OpenRouter, xAI, DeepSeek, Cerebras, Together AI, Perplexity and Ollama Cloud are supported. Paste your key in Settings; it is stored in the macOS Keychain and never sent to us.

Cloud providers are off until you configure one. When you use one, the text for that action goes to that provider under your own account and their privacy policy — see the privacy policy.

Choosing where each feature runs

Settings ▸ AI has two separate choices: the chat provider (the Assistant, and the research behind New Note from a Prompt) and the intelligence provider(Summarise, Suggest Tags/Links, Rewrite, Ask Library and typing suggestions). The intelligence provider defaults to on-device Apple Intelligence. The chat provider is set to OpenAI, but like every cloud provider it stays switched off until you enable it and add a key — so no note text is sent to a model until you choose a cloud provider and set it up. Creativity controls how varied the output is — lower is more focused and predictable.

The same settings say what your chosen model can actually do: whether it runs on your device or in the cloud, roughly how much of a note it can read at once, and which features would work better on a model with more room. An on-device model reads a few thousand characters, so Ask Library over a long note is a different experience there than on a large cloud model — worth knowing before concluding the feature is poor.