Check Knowledge: How to Create a Test from Your Material
The 'Check Knowledge' tool creates a test based on your text: you set the original material, audience and duration, and receive questions with answer options, a separate block of correct answers and explanations. This is not a 'from scratch' generator and not an editor with change highlighting — it is a web form with text fields. In this article we will go through how to fill the fields, how to check the result and where the tool's limits are. All examples are fictional and provided for illustration.
Open tool: Quiz generator ↗Which Fields to Fill and What They Change
"Check Knowledge" accepts three fields: source material (up to 8000 characters), audience or experience level (up to 1000 characters) and length or duration. The output is one result in a separate panel: questions, options, keys and explanations. The output language is selected separately. The result can be copied or downloaded as a plain TXT. There are no style presets, sliders or "level of changes" in the interface — you work with text only.
The first field is not a "topic" but the exact material the test should be based on. If you paste general words like "basics of marketing", the tool cannot rely on specificity and will produce vague questions. Better to give an article fragment, a syllabus, an instruction or a list of definitions. The second field controls the complexity of phrasing and depth: "beginner without experience" and "practicing specialist" will give different tests on the same text. The third field sets the volume: "10 minutes" or "20 questions" — formulate it as you find convenient for checking.
Important: the tool does not verify your material for truthfulness. If the source text contains an error, it may pass into questions and explanations. Therefore the first result is a draft, not a final document.
Source material: "The water cycle includes evaporation, condensation and precipitation. Evaporation occurs from water bodies and soil surfaces. Condensation forms clouds. Precipitation returns water to the ground". Audience: "6th grade students". Duration: "10 minutes".
First Result: What to Check First
After generation open the result and check three things: correspondence of questions to the source text, correctness of keys and quality of explanations. Often the first draft contains questions that test not understanding but memorization of phrasing. For example, a question "Which word in the text stands third in the first sentence?" is technically based on the material but useless for testing knowledge. Such questions need to be reformulated or removed.
Second moment — keys. The tool outputs answers in a separate block, which is convenient: you can take the test first and then check. But make sure the correct option is really one and that it is not duplicated among the incorrect ones. If the material has numbers or dates, check them separately: the model may confuse order or units of measurement.
Third moment — explanations. A good explanation references the source text and helps understand why the answer is correct. If the explanation simply repeats the correct option ("> Correct answer is evaporation, because it is evaporation"), it needs to be rewritten manually or the query clarified. Remember: the model may err, so explanations are a hint, not truth in the last instance.
Draft: "What happens to water during condensation? A) Evaporates B) Forms clouds C) Falls as precipitation D) Freezes". Key: B. Explanation: "Condensation forms clouds". This is correct, but the explanation can be expanded: "During condensation water vapor cools and turns into droplets, from which clouds consist".
How to Refine the Request for a Second Iteration
A second iteration is not a repeat generation "by luck", but an conscious refinement. If the first test turned out too superficial, add more specificity to the audience field: not just "students", but "1st year students, already familiar with basic terms". If questions are too long, specify in the duration field "short questions, 15 items". If explanations are weak, ask to add examples or definitions to the source material — the tool relies on the text, not external knowledge.
It is useful to break the improvement into two passes. First fix the content: remove off‑topic questions, rephrase ambiguous ones, check keys. Then deal with the form: uniformity of options, question order from simple to complex, absence of hints in phrasing. Do not try to do everything at once — you will notice what went wrong faster.
If after a second iteration the result still does not satisfy, check the source material. Perhaps it is too short or contradictory. The tool does not add facts from outside: it works with what you gave. This is its limitation and at the same time its advantage — you control the source.
Was: audience "students". Became: "6th grade students, just read the paragraph, need to check understanding, not memory". Duration: "10 minutes, 8 questions, 4 answer options".
Checklist Before Using the Test
Before giving the test to students or colleagues, take it yourself. Mark questions you hesitated on: if the correct option was not obvious to you, participants will find it even harder. Check that correct answers are distributed across options A, B, C, D roughly equally — if all keys are on "B", this is a hint for guessing.
Verify each question against the source text. If a question requires knowledge not present in the text, it is not a test of the material but a test of luck. If a question allows two correct answers, rephrase it or add "choose the one most accurate". If an explanation contains a statement not in the source, either remove it or add the fact to the source material and regenerate.
Separately check the language: consistency of terms, absence of typos, agreement of cases. The tool may err, and automatic checking does not replace proofreading. If the test will be used for assessment, better save it as TXT and pass it on for review to a person who knows the topic.
Checklist: 1) all questions rely on the text; 2) keys do not duplicate; 3) explanations do not repeat the answer; 4) options are equal length; 5) no questions with two correct answers; 6) taken independently without peeking.
Tool Limitations and Typical Errors
"Check Knowledge" does not export DOCX or PDF, does not do batch processing, does not save results automatically and does not integrate with other systems. You get one result in the panel, which you can copy or download as TXT. This means storage and versioning are your responsibility. Do not expect the tool to remember previous versions: each generation is independent.
The tool does not verify facts and does not search for sources. If the source material has an error, it may end up in the test. The model may err, and queue and availability may change. Therefore do not use the result without proofreading. Also do not count on "level of changes" or highlighting edits — such functions are not in the interface.
Another limitation: volume. The first field — up to 8000 characters, the second — up to 1000. If the material is longer, split it into parts and generate several tests, then combine manually. Do not try to bypass the limit by pasting text in parts into one field — it will not work.
Typical error: a user pastes 12 000 characters into the first field, gets a truncated result and decides the tool "broke". Solution: split the material into two fragments of 6000 characters each and make two tests.
Example of a Full Refinement Cycle
Consider a fictional example. Source material: "The water cycle includes evaporation, condensation and precipitation. Evaporation occurs from water bodies and soil. Condensation forms clouds. Precipitation returns water to the ground". Audience: "6th grade students". Duration: "10 minutes". First result: 6 questions, 2 of which are about word order in a sentence, 1 with two correct answers, explanations short.
Second iteration: refine the audience — "6th grade students, checking understanding of processes, not memory". Duration: "8 questions, 4 options, short formulations". Add to the source material an example: "Over the sea water evaporates, vapor rises up, cools and turns into droplets — so clouds form". Result: questions became more specific, questions on sequence and causal links appeared.
Third iteration: manually correct two questions where ambiguity remains, and rewrite one explanation so it references the example. Check keys: now they are distributed across different options. Save as TXT and pass for review to a teacher. This is the ready material — not perfect, but verified and matching the source text.
Final question: "What happens to water vapor during condensation? A) It returns to the ground as precipitation B) It cools and turns into droplets C) It heats up and rises higher D) It disappears". Key: B. Explanation: "During condensation vapor cools and turns into droplets, from which clouds consist".
Refining a test in "Check Knowledge" is an iterative process: fill the fields, check the first result, refine the request, check again and proofread manually. The tool does not replace an expert and does not guarantee error‑free output, but it saves time on a draft. Remember the limits: no DOCX/PDF export, no batch processing, no automatic saving, and the model may err. Use the source material as the only fact source, check keys and explanations, and the test will become a useful knowledge‑checking tool.
Open tool: Quiz generator ↗Interface translations are machine-assisted and may need refinement.