Filter impressive Claude demo prompts by format, difficulty, audience, and reveal type, then copy the exact prompt and verification checklist.
Disclaimer: the gallery provides reproducible prompt scripts and local tool behavior has been tested. Claude outputs can still vary by model, plan, settings, product surface, and date.
6 demo prompts showing
Artifact
One-Prompt Arcade Game
Beginner10 minutes
Expected result: A playable browser artifact with controls, score, reset state, and a short verification checklist.
<role>You are a playful front-end prototyper.</role>
<goal>Create a single-file Claude artifact for a tiny arcade game.</goal>
<theme>Stage magician catching falling prompt cards before they hit the trapdoor.</theme>
<requirements>
- Ask one clarifying question only if the controls are ambiguous.
- Build the game with keyboard and touch controls.
- Include score, timer, win/lose states, reset, and accessible labels.
- Add a "what to verify before sharing" checklist below the game.
- Keep all assets generated in code; do not use external libraries.
</requirements>
Verification
Play one full round.
Confirm keyboard and touch controls work.
Reset and replay without refreshing.
SVG
SVG Poster Reveal
Beginner8 minutes
Expected result: A polished SVG poster with named layers, editable colors, and a short rationale for the composition.
<role>You are an editorial art director.</role>
<goal>Create an SVG poster that explains a Claude demo as a theatrical reveal.</goal>
<style>High-contrast stage poster, precise labels, no stock imagery.</style>
<requirements>
- Output one complete inline SVG.
- Use named groups for title, background, subject, and annotation layers.
- Include a 5-line rationale after the SVG.
- Make text readable at mobile and desktop sizes.
</requirements>
Verification
Open the SVG preview.
Check that text remains readable.
Change one color variable or group fill.
Data viz
CSV-to-Chart Artifact
Intermediate15 minutes
Expected result: An interactive chart from sample data with assumptions, anomaly notes, and a plain-English takeaway.
<data>
week,channel,spend,qualified_leads
1,Search,1200,18
1,Partner,900,11
2,Search,1350,23
2,Partner,920,14
</data>
<task>Create an interactive visual that shows cost per qualified lead by week and channel.</task>
<requirements>
- First list data-quality checks and assumptions.
- Show the calculation formula.
- Build a reusable chart artifact.
- Include anomaly notes and a "do not overinterpret" warning.
</requirements>
Verification
Recompute one row manually.
Change a sample value and rerun.
Confirm the takeaway updates.
Simulation
Toy Physics Simulation
Intermediate15 minutes
Expected result: A small interactive simulation with sliders, visible assumptions, and a resettable scenario.
<role>You are a science teacher building a safe classroom simulation.</role>
<goal>Make a browser artifact that simulates bouncing balls with adjustable gravity and friction.</goal>
<requirements>
- Include sliders for gravity, friction, and starting speed.
- Show formulas or simplified assumptions in plain English.
- Include pause, reset, and randomize controls.
- Add two questions students should answer after experimenting.
- Avoid claiming the toy model is a real physics engine.
</requirements>
Verification
Move each slider and observe a visible change.
Pause and resume.
Read the assumptions before presenting.
Document
PDF Evidence Table
Intermediate20 minutes
Expected result: A table of claims, page anchors, evidence types, confidence, and manual verification tasks.
<task>Analyze the uploaded PDF as an evidence clerk.</task>
<output>
Return a table with claim, page or section reference, evidence type, confidence, and manual verification step.
</output>
<rules>
- Do not invent page references.
- Distinguish direct text, chart/table, image, and inference.
- End with the three most important unresolved questions.
</rules>
Verification
Spot-check three page references.
Review low-confidence rows.
Keep quotes short and source-bound.
Vision
Screenshot UX Audit
Beginner10 minutes
Expected result: A prioritized UX/accessibility critique grounded in visible evidence from the screenshot.
<context>I am uploading one product screenshot.</context>
<user_goal>The user is trying to choose the right plan and start checkout.</user_goal>
<task>Audit the screenshot for UX, accessibility, and conversion problems.</task>
<output_format>
- Top 5 issues ordered by likely user impact
- Visible evidence, why it matters, suggested fix, confidence
- One thing the screen does well
- One follow-up screenshot to upload next
</output_format>
<constraint>Do not infer hidden flows. Base findings on what is visible.</constraint>