Study tools
These are two free tools I built and continue working on, to help study for a certification exam. Both are prompts, plain blocks of text that you paste or upload into an AI assistant. I keep them downloaded and upload the file fresh at the start of each session. I use Anthropic's models, and Opus set to high effort has worked best for me. I've built and tested both tools on IT certifications, CompTIA A+ Core 2 in particular, but there's nothing tech-specific about them. Any certification with a published set of exam objectives should work.
Both tools generate only original practice material, built from each certification's published objectives. Neither one searches for, references, or reproduces leaked or predicted exam questions.
There's nothing to install and nothing that runs on its own. For each tool below, you can download the .txt version, download the .md version, or open the read view further down to copy the text directly.
Before you start
Both tools work from your certification's objectives document. Download it from the organization that issues the certification, then convert it to markdown to cut down on tokens, or upload the PDF as it is. That's the only setup either tool needs.
The tutor v1.14.0Actively managed: started July 2026. Updated September 2026.
Upload the prompt alongside your objectives list and the tutor starts teaching right away. There are no setup questions and no menu to work through.
The first session opens with a wide sweep to decide what order to teach in. It asks one question for each main numbered objective, using that objective's hardest and most exam-relevant sub-item to stand in for the whole thing. If your objectives list has 60 main numbered objectives, that's a 60 question sweep, not one question per line item. Nothing is scored and you get no feedback, because ordering is the only purpose.
After the sweep, the tutor works one objective at a time: a cold entry test, a teaching pass over whatever you missed, then a closing test over the same ground. It drives the session itself. To skip ahead, slow down, revisit something, or ask for a hands-on lab, interrupt at any point and it follows you.
Verification is strict on purpose. Getting something right during a lesson doesn't count the same as answering a correct question in a verification test, because you just saw the material. An objective is marked verified only when you answer a fresh question later in the session, without help, and explain your reasoning. There's no partial credit. Yes this is intentionally much more strict than most exams.
To finish, type end session. The tutor produces two files: a JSON ledger holding the state of every objective, and a report of what happened. Save both. To pick up where you stopped, upload four files at the start of the next session: this prompt, your objectives list, the newest ledger, and the newest report.
A couple of notes:
- I regularly update these prompts during my study process any time the tools don't behave as intended or I just have a new idea I want to add in. I encourage you to try these out and send me your feedback so I can continue improving them. Just shoot me an email here: [email protected]
- I would recommend completing a first teaching pass through all the objectives in the cert you're studying for with the tutor prompt before starting practice exams. You can keep going with the tutor prompt as long as you like, as its objective is to get you to full mastery on each objective. Realistically you don't need to do this for most certifications though. The exam tool will tell you after your first few practice exams how confident it is in your preparedness without drawing a parallel to the real weighted scores in the actual exams, as it cannot do that. I took 6 practice exams using this tool in the week leading up to taking the CompTIA A+ Core 2 and scored 845 where a pass is a score of 700-900. I believe these tools prepared me adequately and I hope they can assist you with your studies as well.
Read prompt
The practice exam engine v1.4.0Actively managed: started September 2026.
The engine asks three questions up front: which certification you're taking, whether you're attaching objectives, and what you want right now. You can ask for a full-length practice exam, a set scoped to a single domain, or performance-based tasks only. After that it stops asking and gets to work.
It researches the exam before writing a single question. Nothing about the blueprint comes from memory. It verifies domain weights, item count, time limit, question formats, and pass mark against current sources, and it tells you whatever it couldn't confirm. It then allocates question slots to match the published weighting and writes original items. The answer key and rationale are written in the same pass as the question, so nothing gets rationalized after the fact.
What you get back is one self-contained HTML file. Open it in any browser, fully offline, with a countdown timer, review flags, and a confidence tag on every question. Nothing is scored in the page. When you submit, it exports your answers as a markdown file.
To grade the exam, start a fresh chat and upload five files: this prompt, your objectives file, the current ledger, the sealed key for that exam form, and your results export. It grades the exam, writes a debrief you can study from, then writes an updated ledger.
Keep every file it hands you. The ledger is what carries your history between sessions, and without it each test is a sample of one. Tag your confidence honestly too, because a confident wrong answer is the thing worth finding.
Read prompt
Example: performance-based questions
These are three performance-based questions the practice exam engine generated for me while studying for CompTIA A+ Core 2: a network diagram task, a Resource Monitor log task, and a Task Manager task.
Using them together
The two are built to pair up. Learn a domain with the tutor, then sit an exam on it with the practice engine to find out whether it stuck. Both work from the same objectives document.