What Is TokIQ? The Quiz App for Learning Prompt Engineering

TokIQ ("TokIQ: AI Prompt Quiz", by OlaaSoft) is a mobile app for iOS and Android that teaches prompt engineering through short multiple-choice quizzes. Each question is a decision you would make while writing a real prompt: which instruction to add, where a prompt fails, how to fix it. The app is currently in testing and coming soon to the App Store and Google Play.

This page explains what the app does, how it is organized, and who it is meant for, without the marketing gloss.

#Why build a quiz app for prompt engineering?

Most prompt engineering material is written to be read: guides, threads, courses, lists of tips. Reading is useful, and the vendor guides from OpenAI, Anthropic and Google are excellent. But reading a good tip and applying it in the moment are two different skills. Plenty of people can explain few-shot prompting and still write examples that all have the same length, then wonder why every output comes back that length.

Learning research has a consistent answer to that gap: actively retrieving knowledge, by answering questions, makes it stick better than re-reading. We wrote more about this in how to learn prompt engineering. TokIQ applies that idea to one specific skill. Instead of reading about prompting, you make prompting decisions, repeatedly, and get immediate feedback on each.

It is not meant to be the only thing you use. Building small projects and reading vendor documentation still matter. TokIQ is the practice part.

#How does TokIQ work?

The core loop is short, which makes it easy to fit into a few spare minutes.

1. You get a question. It describes a situation from prompt writing and offers several answers. For example, a prompt and its disappointing output, with four possible changes, only one of which addresses the cause.
2. You answer. The app shows right away whether your choice was correct, with a short explanation of the answer. Some questions have more than one correct option and ask you to select all that apply.
3. You can tap "Why?" This opens a short explanation article on the idea behind the question, followed by three follow-up questions on the same concept. The follow-ups are there so that you apply the explanation immediately rather than just reading it.

To make that concrete, here is the kind of decision a question is built around (an illustration, not a verbatim question from the app):

Prompt: "Extract the due date from this invoice email as YYYY-MM-DD."
Problem: For emails with no due date, the model returns a made-up date.

Which change fixes the cause?
A) Add "Be accurate."
B) Add "If no due date is stated, return null. Do not infer one."
C) Ask for the answer in uppercase.
D) Add "You are an expert accountant."

The answer is B, because the prompt never gave the model a legitimate way to report a missing value. The explanation behind a question like this would cover why models fill gaps and why a concrete fallback output works better than general warnings. That pattern (a realistic failure, a few plausible fixes, one that addresses the actual cause) is what most TokIQ questions look like.

The questions and explanations are written with AI and then checked by a separate AI reviewer that has to solve each question without seeing the answer key. Questions that fail the check are not published.

#What topics does TokIQ cover?

There are eight topics, ordered roughly from foundations to advanced system design.

Fundamentals. Clear tasks, context, constraints and output format. The material most everyday prompting problems come down to. See also what is prompt engineering.

Role & context. What a role statement does and doesn't add, what belongs in a system prompt, how much background a model needs.

Examples (few-shot). When examples help, how they get copied too literally, label balance and ordering.

Reasoning. Chain-of-thought, putting reasoning before the answer, and when reasoning models make explicit step-by-step instructions unnecessary.

Structured output. JSON, schemas, field ordering, validation, and handling missing values.

Grounding (RAG). Answering from provided sources, citations, conflicting documents, and saying "not found" instead of guessing.

Tools & agents. Tool descriptions, agent loops, stopping conditions and common failure modes.

Safety (prompt injection). Direct and indirect injection, why delimiters alone are not a defense, and the mitigations that hold up. Background in prompt injection explained.

Each topic has its own page on this site, starting with fundamentals, if you want to see what it covers in more detail.

#What are the levels?

When you set up your profile you choose a level: Beginner, Intermediate or Experienced. Each level serves a different mix of easy, medium and hard questions, and you earn more XP for harder ones.

Beginner questions focus on the everyday decisions anyone using ChatGPT, Claude or Gemini makes. Intermediate questions assume you know the basics and start dealing with trade-offs, such as when examples hurt or how to structure a system prompt. Experienced-level questions move toward the problems people hit when building prompts into software: schema design, retrieval failures, agent loops and injection.

You don't have to start at the bottom if the basics are already familiar. And if you answer almost everything correctly at your current level, the app suggests moving up.

#Can TokIQ adapt to the AI tools I use?

You can tell the app which AI tools you use (ChatGPT, Claude, Gemini, Copilot, Perplexity) and what you mainly use them for (work, study, writing, research or code). That way TokIQ knows the context you are practicing for. The principles behind good prompts are largely the same across these tools, a point covered in prompting ChatGPT vs Claude vs Gemini.

#What else is in the app?

A few practical features, all optional where it makes sense:

• Progress per topic, so you can see which areas you have covered and where you are weaker.
• A day streak and a daily goal (3, 5, 10 or 20 questions) if you want a target.
• A global ranking for registered users, if you like a bit of competition. You appear under a nickname or a masked label, and you can ignore it entirely.
• An evening reminder, off until you turn it on, if you want a nudge to keep a practice habit. Also optional.
• 40 languages for the interface, the questions and the explanations.
• Light and dark mode.

#Is TokIQ free? What's in Premium?

There is a free plan and a Premium plan.

Free:

• 3 topics
• New questions every day in those topics
• 1 explanation ("Why?") per day, plus 1 more by watching an ad
• Includes ads

Premium:

• All 8 topics
• Unlimited questions
• Many more explanations every day
• No ads

The free plan is enough to find out whether this way of practicing works for you. Premium is for people who want the full set of topics, including the more advanced ones, and want to read explanations more often. Pricing will be shown in the app stores at launch.

#Do I need to create an account?

Not at first. You can answer your first two questions as a guest; the app creates an anonymous account in the background. To keep going, you create a free account with email, Google or Apple. It keeps your progress when you change phones and puts you on the ranking. You can delete the account from the app at any time.

#Who is TokIQ for?

People who use AI assistants every day for work, study or writing, and want better results without reading a textbook. The beginner level and the Fundamentals, Role & context and Examples topics cover most of what changes everyday output.

Developers and builders who are putting language models into products. The higher levels and the Structured output, Grounding, Tools & agents and Safety topics are aimed at the decisions that matter when a prompt runs thousands of times without a human checking it.

Students and career changers who want to understand prompting properly rather than collecting templates.

It is probably not for you if you are looking for a library of ready-made prompts to copy. TokIQ deliberately teaches the reasoning behind prompts instead.

#When will TokIQ be available?

TokIQ is in testing now and coming soon to the App Store and Google Play. The download page will have the store links when it launches, and how it works walks through the app in more detail.