Follow the Rabbits: An AI Learning Canvas

Inspired by Alice’s rabbit hole: an AI learning canvas that lets you follow your questions without losing your way. Follow the rabbit, but don’t get lost.

Try it online: https://follow-the-rabbits.simonliu.cn/

Ask AI about an unfamiliar subject, and new ideas keep pulling you along. One term leads to another. Curiosity takes over. Ten minutes later, you are deep down a rabbit hole and have forgotten where you meant to go.

You can swallow the question and stick to the original thread, or follow it and lose that thread altogether. Neither feels right. Curiosity is not the problem. The problem is that curiosity has nowhere to settle. A traditional AI chat window is a one-way street; human curiosity is an underground maze connected to the road above.

Follow the Rabbits, called 寻知问兔 in Chinese and shortened to 问兔, is a place for that curiosity. We keep the Main Path on the surface, give each follow-up question a separate side trail, and mark the rabbit holes you do not have time to enter as Parked. Blind spots and unanswered questions become prompts for a later revisit. When they are due, open Revisit old trails and return with the path you took and the question that brought you there.

01. For People Who Follow Their Curiosity

If learning something new means memorizing the standard answers, Follow the Rabbits may feel unnecessary.

But it may suit you if you recognize yourself here:

You cannot leave an unfamiliar term alone. You read an explanation, get caught by a concept you only half understand, and need to get to the bottom of it before moving on.
You get lost in chat windows. After five or six follow-up questions, you scroll through the conversation and still cannot find the original problem you were trying to solve half an hour earlier.
You want to wander without losing control. You enjoy following a rabbit hole as far as it goes, but want a rope that can pull you back to the surface whenever you need it.

In Follow the Rabbits, you do not have to suppress that wandering. Curiosity is not a distraction from learning. It is your most honest guide to an unfamiliar subject.

02. A Different Job from Mind Maps and Knowledge Graphs

Cards and connecting lines may initially remind you of a mind map or knowledge graph. Start learning with the canvas, though, and the difference becomes clearer.

2.1 Mind Maps Organize What You Understand; This Canvas Helps You Find Your Way Before You Do

A mind map begins with a blank page. It asks you to set out a structure you already hold in your head. When you have just entered an unfamiliar field and do not yet know where its doors are, a blank node offers little help.

Follow the Rabbits starts while you do not yet understand. AI lays out an initial outline of the subject. As you read, anything that puzzles or interests you can become a question attached to that card. You do not have to draw the map yourself. The map records the tracks you leave as you follow the rabbit through the maze.

2.2 A Knowledge Graph Is Someone Else’s Encyclopedia; This Map Records Your Own Detours

A knowledge graph aims for an objective, comprehensive, standardized account. Think of a city transport map hanging on a museum wall: concepts and connections spread across it, impressive to look at, but without telling you where to take your first step.

Follow the Rabbits makes no claim to an all-knowing network. It holds your Main Path and the detours you take from it. Numbered cards show the route toward your goal. A side trail opened by one of your questions hangs below its original card. The map does not aim for an objective or exhaustive picture. It records how you gradually make an unfamiliar subject your own.

03. Start Learning, and Keep a Way Back

There is no complicated environment to configure. Open the site and start looking around.

3.1 Connect Your AI Provider

When you first open Follow the Rabbits, choose a model provider you use and enter your own API key. Learning records stay in this browser. Depending on your choice, the key is stored in this browser or only for the current session. When you call an AI model, the request content and key are sent to your chosen provider; if you enable a proxy, they also pass through that proxy. To look around first, choose Browse first.

3.2 Set a Goal and Lay Out Your First Roadmap

On the empty canvas, describe what you want to learn. Try to specify what you want to be able to do afterward, rather than just naming a broad subject. Before building the roadmap, AI drafts a short learning profile from your description, identifying what you already know and the detours you would like to avoid. Check it, then generate the map. After roughly half a minute—time to get a glass of water—a route through the subject appears on the canvas. Choose a card as your Main Path target, and the cards along the way receive numbers.

3.3 Stay on Your Main Path, and Follow Questions Along the Way

Read the numbered Main Path cards in order. Each card offers a few actions for whatever comes to mind:

ⓘ Explain further: If a card feels too shallow, ask AI to unpack that topic in greater detail, layer by layer.
💬 I have a question: Has a particular term caught your attention? Ask about it directly. The answer grows into a separate side trail immediately below the current card, keeping the earlier context and the Main Path in place.
Mark your understanding: Set a card to Not yet, Mastered, or Blind spot so you can see where your understanding is solid and where it needs work.

3.4 Leave a Marker, Then Return to Your Main Path

When you are deep in a side trail, or there are more interesting questions than you have time for, two features help you keep your bearings:

🔖 Add to Parked: If a side trail will have to wait, choose Add to Parked on its card. This leaves a marker at the rabbit hole and adds it to Parked in the left sidebar. Parked means “keep exploring later.” Return there when you want to pick up the same trail.
Back to Main Path: Whether you have gone ten levels down a rabbit hole or twenty, Back to Main Path brings the viewport back to the nearest point where you left your Main Path.

3.5 Revisit Old Trails

One of the frustrations of learning is understanding something in the moment, then forgetting it three days later. In a card’s details, choose Add to revisits, or use More → Add revisits in bulk at the bottom left. The app prepares questions from your blind spots, unfamiliar concepts, and questions of your own. Preview them and confirm the save. They then appear in Revisit old trails according to their due dates and your daily limit. Each revisit brings back the path you took, the source context, and the question that started it. After you rate your answer, the next revisit is scheduled. Adding a revisit does not mean you have finished exploring that side trail. Its Parked marker stays by default; you decide whether to remove it.

Use Previous / Next to move between items during a revisit session. Your position in the sequence and the number of completed revisits are different things: viewing an answer or moving between items only counts as browsing. An item counts toward the session’s completed revisits only after its rating has been saved successfully. Once the session is complete, you can look back over the items you revisited.

For detailed steps and AI configuration, see the user guide: Main Path, Parked, and Revisit old trails. For how data is handled, see data and privacy. Both reference pages are currently in Chinese.

04. Take Your Learning with You

Every roadmap you build here is yours. Learning records are stored in your local browser. You can export a complete map as a backup, move it to another device, or create a clean Learning Journey page to share with a friend, showing how one question led you all the way down the rabbit hole.

Write down what you want to learn. Follow the rabbit. The way home is still there.