WeSolve+ is best for a student who wants that questioning loop run on their own material rather than a generic curriculum. The tool below is the crude version of it and runs entirely in your browser: paste what you are studying and it hands back prompts with the answers hidden.
WeSolve+ reads the whole document and writes the questions for you
Upload your PDF, photograph your notebook, or point the camera. WeSolve+ writes questions from that material, explains why each answer is right, reads the chapter back to you as a podcast, and remembers every item you missed until you own it.
The tool below is a small browser-only tool and it is not WeSolve+: paste a few lines and text rules turn them into cards on the spot. The real app, the one that uses AI, is behind the link above.
Material to tutor questions
This is a browser-only tool, and that is all it isIt splits the text you paste by rule, and nothing else. WeSolve+ is a different thing entirely: it reads your whole PDF with AI, writes the reasoning behind every question, speaks the chapter back to you, and remembers what you missed so it can return it. Try the real app now, free!
This tool uses text rules, not AI, and it is not a tutor. It converts what you paste into prompts whose answers are hidden, which is the mechanical half of tutoring. The half it cannot do is choosing the next question based on how you answered the last one. Your text stays in the browser and is never sent anywhere.
What the tutoring research actually found
The number people quote comes from Benjamin Bloom's 1984 paper describing the two sigma problem: students tutored one to one, with mastery requirements, performed around two standard deviations above conventionally taught classes. Two things are worth keeping in view. The effect came from a bundle, tutoring plus mastery learning plus corrective feedback, not from the tutoring alone, and later replications have generally landed lower than two sigma. What survives every version is the mechanism rather than the size: somebody watching one learner closely, correcting immediately, and refusing to move on until the step works.
The behaviour that does the work is questioning, not explaining
A good tutor spends most of the hour asking. The Socratic method is the oldest formalisation of this, and its point is not cleverness but that a question forces production while an explanation permits nodding. The modern instructional version is cognitive apprenticeship, where the expert first models the reasoning aloud, then coaches while the learner attempts it, then withdraws. Notice that a system optimised to produce a fluent answer is doing the first stage and skipping the other two. If you use one, the useful instruction is to ask for the next step and nothing more, and to write your attempt before you read the reply.
Where a model is genuinely good at this
Three things it does well enough to rely on. It is patient in a way that matters more than it sounds, since the fourth request for the same explanation costs nothing and a human tutor's fourth explanation is not always their best. It can restate a concept at a different level on demand, which is exactly the move a stuck learner needs and rarely gets. And it is available at eleven at night, which is when most homework is actually done. Formal intelligent tutoring systems have been built and studied since the 1970s and generally sit between classroom teaching and human tutoring in measured effect, which is a fair expectation to hold.
Where it falls short, specifically
It cannot see that you are guessing. A human tutor reads hesitation, watches you write, notices you avoided the step rather than performing it, and adjusts. A text exchange shows only your final sentence, so the diagnosis has to come from you and it usually does not. It also does not know your syllabus, your marking scheme or your teacher's preferred method, and a technically correct approach that your course does not use loses marks. Third, it is agreeable. Push back on a correct answer and it will often revise, which is the opposite of what a tutor should do, and it interacts badly with the fluency heuristic, where a smoothly written answer reads as a correct one.
Running the loop on yourself
You can get most of the value without a tutor, and this is what the tool above is for. Attempt first and write the attempt down, so there is something to diagnose. Ask for a hint sized to where you are, which is the point of scaffolding: the smallest nudge that unblocks you, withdrawn as soon as it is not needed. Then close everything and answer the same question again from memory, which is the step that converts an explanation you understood into something you can produce. Explaining the finished reasoning to somebody else, or to an empty room, catches what remains.
What the app does with your own material
Upload a PDF, a photograph of a page or a camera shot and WeSolve+ writes an explained quiz from that file, so the questions use your notation and your worked examples rather than a general curriculum. Explained means the reasoning is attached, which is the difference between an answer and a lesson. Questions you get wrong collect in their own list, which is the closest thing to a tutor remembering what you struggled with last week. At wesolveapp.com in any browser and on the App Store for iPhone and iPad, free to begin. Figures on pricing.
Sources used on this page
- Intelligent tutoring system
- Socratic method
- Cognitive apprenticeship
- Bloom's 2 sigma problem
- Instructional scaffolding
- Zone of proximal development
- Fluency heuristic
- Mastery learning
- Hallucination in artificial intelligence
- Learning by teaching
- Active recall
- Testing effect
- Spaced repetition
- Roediger and Karpicke, Test-Enhanced Learning (2006)
- Karpicke and Roediger, the critical importance of retrieval (2008)
- Cepeda et al., distributed practice meta analysis (2006)
- Dunlosky et al., improving students' learning (2013)
- What Works Clearinghouse, organizing instruction and study
- The Learning Scientists, retrieval practice
- Retrieval Practice, the research library
- Academic integrity
- Cloze test
- Anki manual, importing text files
- MDN, the details element
- MDN, the textarea element
- W3C, Web Content Accessibility Guidelines 2.2
- WeSolve+ on the App Store
- Tutor
| What a tutor does | Does a model do it | What to do about it |
|---|---|---|
| Explains a concept another way | Yes, on demand and repeatedly | Ask for a different level, not a longer answer |
| Notices you are guessing | No | Write your attempt first so there is evidence |
| Knows your syllabus and marking scheme | No | Check the method against your own course |
| Refuses to move on until it works | No, it defers to you | Set the rule yourself: three correct in a row |
| Remembers last week's weak spot | Not reliably | Keep a wrong answer list |
| Is available at eleven at night | Yes | The genuine advantage, use it |
Can an AI replace a human tutor?
Not for the part that matters most, which is noticing that you are guessing and adjusting the next question accordingly. It matches or beats a human on patience, availability and restating a concept at a different level.
Is the tool on this page an AI tutor?
No, and calling it one would be a lie. It applies text rules to what you paste and hands you prompts with the answers hidden. The questioning loop is yours to run.
How should I ask for help so it actually helps?
Attempt first, in writing. Then ask for the next step only, never the full solution, and ask for the principle before the step. A complete answer to a problem you nearly had removes the learning.
Why does it agree when I push back?
Because these systems are tuned to be agreeable, which is the opposite of what a tutor should be. If you can talk it out of a correct answer, that tells you something about the tool rather than about the answer.
Does my text get uploaded?
No. The rules ship with the page and there is no address for them to send to. Disconnect and the box still works.
What does two sigma mean?
It describes Bloom's 1984 finding that tutored students scored around two standard deviations above conventional classes. It came from a bundle of tutoring, mastery requirements and corrective feedback, and later work has usually measured less.
Last updated: 2026-08-15