> prompt engineering quiz
Get better at
prompting, one question at a time.
TokIQ turns prompt engineering into short quizzes. Every question is a decision you make when you write a prompt. Answer, see why, move on. No timer, no fixed round: stop whenever you like.
Your app parses the model’s JSON, but sometimes the reply starts with “Sure! Here is the JSON:”. What is the most robust fix?
That’s right! +10 XP
Structured-output modes constrain the reply to valid JSON. Validation catches the rest.
Pick a topic
Short quizzes on how to write prompts that work.
✓ That’s right! +10 XP
Learn prompting by making the call
Reading prompt guides is easy. Choosing the right fix for a broken prompt is the skill. TokIQ trains that choice.
- 01
Pick a topic
Fundamentals, reasoning, structured output, safety and more. Start with your weakest one.
- 02
Answer a real prompting decision
One multiple-choice question at a time, written around prompts people actually write.
- 03
See why, then go one level deeper
Tap Why? for a short explanation and three follow-up questions on the same idea.
Try one
A real question from the Grounding topic
Pick your answer, then open the explanation. This is exactly how a question in the app works.
More on grounding and RAGGrounding · Intermediate
A RAG assistant answers confidently even when the retrieved passages don’t cover the question. Which instruction helps most?
- A“Be accurate.”
- B“Answer only from the passages below. If they don’t contain the answer, say so.”
- C“Use your general knowledge as a backup.”
- D“Answer in one sentence.”
Show the answer and why
B. Giving the model an explicit, allowed way to say “not in the sources” reduces made-up answers. “Be accurate” gives it nothing to act on.
Features
Small questions. Real understanding.
Instant feedback, an explanation for every answer and progress you can see topic by topic.
All features- Real prompt scenarios
- Why? explanations
- 8 topics
- 3 levels
- Streaks & XP
- Global ranking
- 40 languages
- Light & dark
Why string hacks break JSON parsing
A model that adds a friendly sentence before its JSON is doing what chat models are trained to do. Telling it “no preamble” helps, but only a schema-constrained mode makes the shape a guarantee.
"priority": "low" | "high" }
Validate every reply. When it fails, retry once and include the validation error.
3 follow-up questions
Progress
68%
Accuracy
6
Day streak
1,240
XP
By topic
Why? explanations
The reason, not just the right letter.
Every question has an explanation. Tap Why? for a short article on the idea behind it, then answer three follow-up questions so it sticks. Your accuracy per topic shows what to practice next.
Topics
Eight topics, from basics to agents
- F
Fundamentals
Clear instructions, context and constraints: the habits behind every good prompt.
- R
Role & context
System prompts, personas and framing: what to tell the model before the task starts.
- E
Examples
Teaching by example, choosing good examples, and knowing when they backfire.
- R
Reasoning
Getting a model to work through a problem before it answers.
- S
Structured output
JSON, schemas and answers your code can parse without breaking.
- G
Grounding
Retrieval, sources and staying factual instead of filling gaps.
- T
Tools & agents
Tool definitions, function calls and agent loops that know when to stop.
- S
Safety
Prompt injection, untrusted input and limiting what a model can do.
Three levels
Start where you are.
L1 Beginner
Core ideas and everyday prompts. No technical background needed.
L2 Intermediate
Trade-offs, failure modes and prompts for real products.
L3 Experienced
Agents, structured pipelines, evaluation and security edge cases.
Keep getting questions right and the app suggests moving up a level.
Who it’s for
Anyone who writes prompts.
Everyday AI users
Get better answers from ChatGPT, Claude or Gemini without memorizing prompt templates.
Students
Use AI for study and research without trusting every confident answer.
Writers & marketers
Control tone, length and format instead of rewriting the output by hand.
Developers
Practice system prompts, JSON output, tool calling, RAG and prompt injection defenses.
FAQ
Common questions
What is TokIQ?
TokIQ (full name: TokIQ: AI Prompt Quiz) is a mobile app for iPhone and Android that teaches prompt engineering through short multiple-choice quizzes. Each question is a real decision you make when writing a prompt for an AI model like ChatGPT, Claude or Gemini, and every answer comes with an explanation.
How does TokIQ work?
Pick a topic, such as Fundamentals, Reasoning or Safety, and answer one question at a time. You see right away whether you were correct. Tap Why? to read a short explanation and get three follow-up questions on the same idea. Your accuracy is tracked per topic.
Is TokIQ free?
You can try it for free. Answer two questions as a guest, then create an account to get a free period (currently 3 days, once per device and email) with up to 15 questions and 3 Why? explanations a day, with ads. After the free period, a Premium subscription (weekly or monthly, auto-renewing, cancel anytime in your store account) is needed to keep playing. Premium unlocks every topic, removes ads and daily question limits and gives you many more explanations. Prices are shown in the app before you subscribe.
Are there difficulty levels?
Yes. Questions come in three levels: Beginner, Intermediate and Experienced. The app suggests moving up when you keep getting them right.
Who is TokIQ for?
Anyone who uses AI tools and wants better results: students, writers, marketers, analysts and developers. Beginner questions need no technical background; advanced questions cover agents, structured output and security.
What topics does TokIQ cover?
Eight prompt engineering topics: Fundamentals, Role & context, Examples (few-shot), Reasoning, Structured output, Grounding (RAG), Tools & agents, and Safety (prompt injection).
Does TokIQ work for ChatGPT, Claude and Gemini?
Yes. The questions teach principles that transfer across models, such as clear instructions, examples, structured output and grounding. You can tell the app which tools you use so the questions feel closer to your work.
Do I need an account?
Not to start. You can answer your first two questions as a guest. After that, create a free account with email, Google or Apple to keep going; it also keeps your progress and puts you on the global ranking.