Learning Playground

Follow a thought down the rabbit hole. A Playground is a canvas where you explore a topic and your local AI deals you knowledge cards — short, rich mini-articles it writes on the spot — one leading to the next, as far as your curiosity runs. The thinking is done on your Mac. Pictures are the exception, and only when you ask for them: those are searched for in openly-licensed collections, using the words you typed and nothing else from your wiki.

Start exploring

Choose ✦ New Playground (from the ⇧⌘P command palette, or the New menu) and type what you want to explore — a concept, a question, a name. MojoPad lays down a seed and deals your first card: a written answer with headings, key terms in bold, and (if you like) a photo.

To go deeper, press ✦ Grow on any card and a fresh card is dealt from it — the exploration branches like a real train of thought. Every card also has an “Add a follow-up…” bar at its foot: ask a pointed question and the answer arrives as its own card. Bold keywords are clickable — click one to grow a card about exactly that.

Reading the board

• Trails color-code themselves — each line of inquiry takes a color, so you can see at a glance which branch is which.
• The frontier — the newest, least-explored edge of the board carries a soft amber tint and a gentle nudge to keep a card worth keeping.
• The view follows you — a Grow never changes your zoom; the board simply pans to reveal the new card, so you never lose your place.
• Cards size to their content — a card is never a tiny box with a scrollbar; it grows to fit what it says.

Photos

Turn on Photos and cards illustrate themselves with openly-licensed images (from Openverse and Wikimedia Commons), chosen for relevance to the subject — the mechanism, not just the word. Click a photo for a full-size lightbox with its attribution and license.

It's still a page

A Playground is a Canvas underneath, so it saves like any other page, turns up in search, and is there to revisit or keep growing whenever you come back. Drag a card off to a page, or fold the whole thing into your notes. And because it runs on your local model (see Local AI), you can explore freely — no accounts, no tokens, no limits.

When this Mac hasn't the memory for it

The cards are only ever as good as the model writing them, and a small model writes thin, general ones. That is the honest limit of running everything here: on a Mac without the memory for a capable model, the Playground is the one part of MojoPad you cannot improve by using it better.

So if you already have your own model set up — Settings ▸ AI ▸ Send the questions I ask to a model I provide — there is a second box beneath it: Let the Playground use it too. Off unless you tick it, and worth understanding before you do:

• Your cards then come from that model, and the Playground works properly on a machine that cannot run one locally.
• You ask a Playground by clicking, and a good rabbit hole is twenty cards in ten minutes. Each one is billed by your provider, where a drawer question was one sentence you sat down to write. That difference is the whole reason this is a separate box rather than part of the first one.
• Permission is still per wiki. A wiki you have never allowed to send still cannot, so its Playground carries on writing here — which means one board can hold cards of both kinds.
• Every card says which model wrote it. The chip on an AI card is marked ⌂ when it was written on this Mac and ↗ when it came from your own model, and its menu says so in words. That is taken from what actually happened to that card, not from what the setting says today.

Leave the box alone and nothing changes: the Playground is local, as it always has been.