The Muse Spark prompt generator gives you 20 free, copy-ready prompts for Meta's Muse Spark AI model. Generate interactive web apps, data dashboards, long-form writing, and code — from a single natural language prompt. Available across WhatsApp, Instagram, Facebook, and Meta.ai.
The Muse Spark prompt generator on this page provides 20 free, professionally crafted prompts for Muse Spark, the AI model developed by Meta Superintelligence Labs and released in April 2026. Muse Spark is the first model from Meta's new research division — and its defining capability is visual coding: generating functional, interactive web applications and dashboards from natural language descriptions.
Unlike models that output code you then have to run yourself, Muse Spark renders working interactive tools — habit trackers, data calculators, quiz apps, expense dashboards — directly in the Meta AI interface. It also accepts voice and image inputs alongside text, and produces high-quality long-form writing across formats. With distribution across WhatsApp, Instagram, and Facebook, it reaches more users than any other AI model released in 2026.
The 20 prompts below are split between Muse Spark's two core capabilities: visual coding (apps and dashboards) and writing (articles, scripts, emails, social content). Every prompt is structured to take advantage of Muse Spark's specificity preference — detailed prompts produce significantly better outputs than vague ones.
Click any prompt to copy — paste into Meta AI on meta.ai, WhatsApp, Instagram, or Facebook
Build an interactive web dashboard showing global renewable energy adoption from 2000 to 2026. Include: a line chart comparing solar, wind, and hydro capacity growth by continent; a choropleth map showing current renewable percentage by country; a summary card showing total global renewable capacity in terawatts. Use clean, modern design with dark mode. Data should be realistic and sourced to Our World in Data conventions. Dashboard must be fully interactive — clicking a country on the map filters the line chart to that region.
Write a 600-word explainer article on how transformer attention mechanisms work, aimed at a software developer who understands Python but has no ML background. Start with an analogy that makes the concept intuitive. Then explain: (1) what a token is; (2) what attention scores represent; (3) why multi-head attention is used; (4) what the output actually is. Include one simple pseudocode block showing scaled dot-product attention. End with one sentence on why this matters for language models. Tone: clear, confident, no unnecessary jargon.
Create a functional habit tracker web app. Features: add up to 10 habits with custom names; check off habits each day; show a 30-day streak calendar per habit (green = completed, grey = missed, white = future); display a summary score for today's completion percentage; persist state in localStorage so habits survive page refresh. Design: minimal, satisfying to use, mobile-responsive. No external libraries — vanilla JS only. The streak calendar should be the visual centrepiece.
Write a 5-minute podcast script in interview format for a show called 'Founders at Work.' The guest is a first-time founder who just closed a seed round for a climate tech startup that captures methane from landfills. The host asks: (1) what made them leave a stable job to start this; (2) what the technology actually does; (3) what they got wrong in the first 6 months; (4) what surprised them about fundraising. The founder's answers should feel authentic — specific, slightly imperfect, with personality. Not corporate. Include natural transitions and a closing question about advice for first-time founders.
Build a compound interest calculator web app that shows wealth growth over time. Inputs: initial investment, monthly contribution, annual interest rate, investment period in years, tax rate on gains. Output: (1) an animated line chart showing portfolio value year-by-year; (2) a breakdown table showing contributions vs. interest earned vs. tax paid per decade; (3) a summary showing final value, total contributed, and total interest earned. The chart should animate on first load. Design should be clean, professional, usable for financial planning conversations.
Write a 700-word science fiction short story set in 2089. The premise: a climate engineer is responsible for managing rainfall patterns across Central Asia using atmospheric modification technology. The story takes place on the day she discovers that her algorithm has been subtly adjusted by an unknown party — and the adjustment, if left running, will redirect rainfall away from 40 million people over the next decade. She has 6 hours before the next atmospheric injection cycle locks the change in. The story should end ambiguously. No exposition dumps. Start in scene.
Design and build the popup UI for a browser extension called 'Focus Guard.' The extension blocks distracting websites during focus sessions. The popup should show: a large timer display for the current session (MM:SS format); a start/pause/stop button; a list of blocked sites the user has added; an 'Add site' input field; a toggle for 'break mode' (5-minute timer, all sites unblocked). The timer counts down from a user-set duration (default 25 minutes). Popup dimensions: 320px × 480px. Design: dark theme, orange accent color, feels like a focus tool not a productivity app.
Write a 500-word argumentative essay making the case that four-day work weeks increase long-term productivity in knowledge work roles. Structure: a hook opening with a counterintuitive statistic; a thesis; three evidence-based arguments (cite specific studies or real-world examples from Iceland, Microsoft Japan, or similar); one counterargument acknowledged and rebutted; a closing that connects to the broader question of what 'productivity' means. Tone: persuasive but not polemical. Suitable for a business publication.
Build a 10-question multiple-choice quiz app on world geography. Each question has 4 answer options. Features: one question shown at a time with a progress bar; immediate feedback on answer selection (green = correct, red = wrong) with a one-sentence explanation; a final score screen showing percentage correct and a performance label ('Expert,' 'Traveller,' 'Beginner'); a retake button that randomises question order. Design: clean cards, no clutter, mobile-first. Questions should cover capitals, rivers, mountain ranges, and borders across all continents.
Write a 4-email onboarding sequence for a project management SaaS tool called Frameflow. The sequence runs over 7 days post-signup. Email 1 (day 0): welcome + one action to complete in the next 10 minutes. Email 2 (day 2): spotlight one feature (task dependencies) with a use-case example. Email 3 (day 4): social proof — one customer outcome story in 3 sentences + a feature the customer used. Email 4 (day 7): 'check-in' — ask if they've hit any friction + point to help resources. Tone across all 4: warm, direct, no corporate fluff. Subject lines included.
Build a local weather dashboard web app. Use the Open-Meteo API (no API key required) to fetch weather data. Features: auto-detect user location via browser geolocation; display current temperature, weather condition, wind speed, humidity, and UV index; show an hourly forecast for the next 12 hours as a scrollable card row; show a 7-day forecast with high/low temperatures and condition icons. Design: clean and airy, inspired by Apple Weather — use weather-appropriate background gradients (blue for clear, grey for cloud, dark for night). Fully responsive.
Write three different product descriptions for the same item: a handmade ceramic coffee mug, 350ml, matte white with a speckled glaze, made by a small studio in Portugal. Version 1: Amazon listing style — benefit-focused, scannable bullet points, SEO-optimized for 'handmade ceramic mug.' Version 2: Direct-to-consumer brand style — story-forward, emotional, 80 words, makes the buyer feel something. Version 3: Wholesale catalogue style — functional, technical, precise measurements, no marketing language. All three should be immediately usable without editing.
Build a minimal Markdown note-taking app that runs entirely in the browser. Features: a sidebar listing all notes with titles extracted from the first H1; a split-pane view with Markdown source on the left and live-rendered HTML on the right; create, rename, and delete notes; auto-save to localStorage every 2 seconds; keyboard shortcut Ctrl+N for new note. Design: two-pane layout, monospace font in editor, comfortable line height in preview, dark mode by default. No external dependencies except a lightweight Markdown parser (marked.js or similar via CDN).
Write 3 different LinkedIn posts on the topic of 'what separates good managers from great ones,' each with a different angle and format. Post 1: Personal story format — a specific moment where a manager changed your perspective, ending with a broader lesson. Post 2: Contrarian take — challenge a common piece of management advice and explain why it backfires. Post 3: List format — '5 things great managers do that good managers don't,' where each item is specific enough to be actionable, not just aspirational. Each post should be under 220 words. Include first lines designed to stop the scroll.
Build a personal expense tracker web app. Features: add expenses with amount, category (Food, Transport, Housing, Entertainment, Health, Other), date, and optional note; display a running monthly total; show a donut chart breaking down spending by category; list all transactions in reverse chronological order with delete option; show a simple insight: which category is highest this month vs. last month. Persist all data in localStorage. Design: clean, minimal, feels like a personal finance tool. Monthly navigation (previous/next month buttons).
Write three versions of a speaker bio for Dr. Amara Osei, a Ghanaian-British economist and author who advises governments on digital infrastructure policy. She has a PhD from LSE, has published two books, and speaks at Davos, TED, and UN events. Version 1: Conference programme bio — 60 words, third person, credential-focused. Version 2: Event website bio — 120 words, third person, narrative flow, ends with why she's relevant to the audience. Version 3: First-person bio for her own website — warm, direct, explains her work in plain language, mentions her background without it reading like a CV.
Build a Pomodoro timer web app. Features: 25-minute work session timer; 5-minute short break; 15-minute long break after every 4 pomodoros; a session counter showing how many pomodoros completed today; a subtle sound notification when a timer ends (use the Web Audio API to generate a gentle tone — no external audio files); auto-start next phase option toggle; session history log showing times of completed pomodoros. Design: centred, large timer display, muted colour palette that doesn't compete with work — soft cream background, dark text, one accent colour for the active state.
Write a weekly newsletter issue for a publication called 'The Signal' covering emerging technology for a non-technical professional audience. This week's issue: theme is 'the attention economy is breaking.' Include: (1) a 150-word opening essay making an original argument about how notification design shapes decision-making; (2) three curated links with 30-word summaries explaining why each matters; (3) one 'counterintuitive thing I learned this week' — a specific fact or observation, not a platitude; (4) a closing one-liner that leaves the reader with something to think about. Tone: intelligent, opinionated, not trying to be liked.
Build a recipe finder web app using the MealDB API (free, no key required). Features: search recipes by ingredient or name; display results as a card grid with photo, recipe name, and cuisine type; click a card to open a modal showing full recipe: ingredients list, step-by-step instructions, and a YouTube embed if available; a 'random recipe' button that fetches a surprise meal; save up to 10 favourites in localStorage with a star toggle on each card. Design: warm, food-appropriate — use earthy tones, generous imagery, comfortable reading layout for instructions.
Write a cover letter for someone making a career change from secondary school teaching (10 years, English and History) into UX research. They have completed a UX research bootcamp, conducted two portfolio case studies (one on a school app, one on a local transit service), and have a genuine interest in research methodology. The letter is for a Junior UX Researcher role at a fintech company. It should: acknowledge the career change directly rather than avoiding it; frame teaching skills (observation, facilitation, synthesis of qualitative feedback) as directly transferable; reference the portfolio work specifically; be 300 words, no fluff, no clichés ('I am passionate about...').
Muse Spark's visual coding and Meta distribution are its key differentiators:
| Model | Visual Coding | Voice Input | Distribution |
|---|---|---|---|
| Muse Spark (Meta) ★ | Yes — renders interactive apps | Yes — native | WhatsApp, Instagram, Facebook, meta.ai |
| GPT-5.5 (OpenAI) | Code output (not rendered) | Yes — Advanced Voice | ChatGPT web, API |
| Gemini 4 (Google) | Code output (not rendered) | Yes — Live mode | Gemini app, Google Search, Meet |
| Claude (Anthropic) | Code output via Artifacts | No | claude.ai, API |
★ Muse Spark released April 2026 by Meta Superintelligence Labs. Available free across Meta's apps with 3+ billion combined monthly users.
The Muse Spark prompt generator on this page provides 20 free, copy-ready prompts for Meta's Muse Spark AI model, developed by Meta Superintelligence Labs and released in April 2026. Muse Spark accepts voice, text, and image inputs and generates not only text but also interactive web applications, data dashboards, and visual tools in a single output. The prompts on this page cover both its writing and visual coding capabilities.
Muse Spark is the first model released by Meta Superintelligence Labs (MSL), the AI research division Meta formed in early 2026 by merging and expanding its existing AI teams. Unlike Meta's previous Llama line of open-weight models, Muse Spark is proprietary and not open-weight. It is distributed through Meta AI across WhatsApp, Instagram, Facebook, and Meta.ai on the web. Muse Spark's defining capability is 'visual coding' — generating functional interactive web applications and dashboards from natural language prompts, in addition to high-quality long-form writing.
Muse Spark's primary differentiation is native visual coding output: it generates working interactive web apps and dashboards — charts, calculators, trackers, quiz tools — directly from a text description, without requiring the user to have coding knowledge or a separate tool. ChatGPT and Gemini can generate code, but Muse Spark generates rendered, functional applications as a first-class output type. It also accepts voice input natively across Meta's apps, and processes image inputs for visual tasks. The combination of massive distribution (Meta's apps have 3+ billion users) and visual coding capability is unique.
Muse Spark responds well to structured, specific prompts. For visual coding (apps/dashboards): describe the exact features you need, the data sources to use (prefer free APIs like Open-Meteo or MealDB for live data), the design style in concrete terms, and any interaction behaviour. For writing: specify format, word count, tone, target audience, and any specific constraints or examples. Muse Spark handles long detailed prompts better than vague short ones — more specification produces better output. For interactive apps, list every feature explicitly rather than leaving it to interpretation.
Muse Spark is available through Meta AI at meta.ai, within WhatsApp (via the AI chat interface), on Instagram (via the AI button in DMs and in the search bar), and within Facebook. Meta AI is free for all users on these platforms. Enterprise and API access is available through the Meta for Business portal. As of May 2026, Muse Spark has no standalone app — access is entirely through Meta's existing platform distribution.
Yes — Muse Spark's visual coding capability generates functional HTML/CSS/JavaScript applications that run in the browser. These are not mockups or wireframes but working interactive tools: a habit tracker that saves to localStorage, a compound interest calculator with animated charts, a quiz app with scoring logic. The output is a complete, deployable web application you can copy and host. For data-heavy apps, Muse Spark integrates with public APIs (weather, food, finance) natively when they are specified in the prompt.
OpenAI's default ChatGPT model — writing, coding, research
Google's flagship AI — reasoning, multimodal, coding
xAI's FLUX.1-powered image generator
OpenAI's latest image model — portraits, products, posters
Google's best-in-class photorealistic image generation
Build structured prompts for any AI image model