Muse AI Cheat Sheet: Prompts, Features & Settings

The 5 prompt formulas

1. Outcome first: “Book me a flight to Chicago next Friday morning, under $300, aisle seat.” Muse plans backward from a clear goal.
2. Context sandwich: who it's for + constraints + deadline. “Vegetarian dinner for 6, under $80, delivered by 7pm.”
3. Plan before acting: “Outline how you'd research this before you start.” Catches misunderstandings when they're free.
4. Approval checkpoint: “Find three options, then stop and let me pick before booking.” The highest-leverage habit in agent work.
5. Iterate, don't restart: “Make it shorter,” “add prices,” “sort by rating.” Agents handle revision rounds gracefully.

Features at a glance

Every major Muse feature, what it does, and when to reach for it:

Settings worth changing today

• Approvals: keep human review ON for sending messages, purchases, and file changes — the whole safety model is five seconds of your attention.
• Notifications: allow approval pings so Muse can reach you when it's waiting on a decision; mute everything else.
• Profile preferences: favorite airline, dietary needs, budget ranges, how you like summaries formatted — teach once, benefit forever.
• Memory controls: review what Muse remembers about you; tighten or loosen to taste.
• Voice: enable voice mode for brainstorming and hands-free moments; keep text for anything that must be quoted exactly.

What Muse can't do

Honest limits, so you don't learn them the embarrassing way:

• Read your mind — every missing constraint (budget, date, audience) is a guess it has to make.
• Guarantee facts — verify prices, dates, and recipients before acting; polished output isn't proof.
• Reach services you never connected — no connector, no real data, just general knowledge.
• Undo the irreversible — sent emails and completed purchases are why approval checkpoints exist.
• Work outside its availability — features, limits, and access vary by region and account.

Quick answers

Do I need to learn prompt engineering? No — the five formulas above cover nearly everything. Agent prompting is mostly saying what you want, what it needs to know, and where to pause.

What's the one habit that matters most? Approval checkpoints. Tell Muse where to stop and wait for you, especially before money moves or messages send.

When should I use voice vs. typing? Voice for exploring and brainstorming; typing for addresses, numbers, and anything quoted exactly. Details: voice mode guide.

Where do I go deeper? The prompt patterns guide expands these formulas into ten full patterns, and the tutorial walks through your first real tasks.

Keep this bookmarked

Print it, screenshot it, or just remember the one-liner: outcome first, context second, checkpoint third. When you're ready for the full vocabulary behind these features, the glossary defines every term plainly — and the 50-things list gives you fifty jobs to try this week.