Games

Games · Prompt Arcade

Prompt engineering, played.

36 short games that train how you prompt. Six hands-on rounds each — no reading lists, no guessing the longest answer. Every round shows you exactly why.

Daily challenge5 rounds · new every dayMixed from the free games. Beat your score, share it.Play →
Spot it

Tap the lines that will make the AI fail — or the claims the source doesn't support.

Build it

Pick a card for each slot, then see exactly what the AI does with your choices.

Match it

Link each prompt to the answer it produced. Learn cause and effect.

Order it

Put the steps of a prompt chain or agent loop in the right order.

Start here

5 games · Free — no account needed.

Quick wins

8 games · Five-minute fixes you'll use today.

Advanced techniques

12 games · Reasoning, grounding, agents, evaluation.
AdvancedNewSelf-Consistency VotingRun the same prompt 3-5 times → take the majority answer. +12-18% accuracy on reasoning tasks (Wang et al. 2022).Spot itOrder itMatch itBuild itFree with an accountAdvancedNewFew-Shot Pattern RecognitionShow 2-4 examples → the model maps the relational structure, not the surface details (Chase & Simon's chess masters, applied).Build itSpot itMatch itOrder itFree with an accountAdvancedNewConstraint CalibrationToo few constraints → drift. Too many → brittle. The 5-constraint sweet spot and how to find it.Spot itBuild itMatch itOrder itFree with an accountAdvancedNewMulti-turn RefinementConversation as compute. Use 3-5 turn loops when the goal is discovered, not pre-specified.Order itSpot itBuild itMatch itFree with an accountAdvancedNewMeta-PromptingAsk the model to write its own prompt. The technique that compresses 4 hours of prompt-engineering into 4 minutes.Order itBuild itSpot itMatch itFree with an accountAdvancedNewPrompt Evaluation RubricsBuild your own grader. The Chip-Huyen-style eval rubric that catches drift before it ships.Build itSpot itOrder itMatch itFree with an accountAdvancedNewAdvanced Chain-of-Thought MasteryStack CoT with role + self-critique for 60%+ accuracy on multi-evidence reasoning.Order itSpot itBuild itMatch itFree with an accountAdvancedNewRole & Persona Prompting That Actually WorksMulti-dimensional personas that shape voice, taste, and decision-making — not a frankenstein average of 10,000 examples.Build itSpot itMatch itOrder itFree with an accountAdvancedNewOutput Formatting SecretsSchema enforcement, prefilling, XML tags — the production tricks vendors don't teach.Build itSpot itMatch itOrder itFree with an accountAdvancedNewCut Hallucinations: Grounding, Sources, UNKNOWNProduction-grade hallucination prevention via citation-or-refuse + admit-uncertainty + claim triage.Spot itBuild itMatch itOrder itFree with an accountExpertNewBuilding Simple Agents (ReAct)Thought → Action → Observation — the foundation of every modern agent, with halt conditions that don't burn $500.Order itSpot itBuild itMatch itFree with an accountExpertNewTree-of-Thoughts for PlanningWhen linear reasoning fails — enumerate branches, score independently, surface trade-offs, recommend a hybrid path.Order itSpot itBuild itMatch itFree with an account

Role spotlights

6 games · Prompting for legal, SaaS, founders, marketing, content.

Mastery

5 games · Put it all together.

Why games work for prompting

Prompting is a skill, not a fact list. You get better by spotting what breaks a prompt, rebuilding it, and seeing the effect — so every round asks you to do exactly that, then explains the pattern behind the answer. Scores are saved in your browser, and on your account when you're signed in. Want ready-made prompts instead? Browse the packs or the free prompts.