Day 1
What Lovable actually is
Understand the model: you describe intent in chat, an agent writes real code, and the preview shows the result live.
Concepts
- · Lovable generates a genuine React codebase — not a locked no-code canvas
- · Chat mode discusses and plans; build mode edits files
- · The preview is your app running live; every message can change it
- · Free vs paid: credits per month, private projects, custom domains and higher limits on paid plans
Exercise
Start a project with a one-paragraph description of something small but real from your own work — a form, a tracker, a directory. Don't fix anything yet; just read what it built.
Try this prompt
Build a [thing] for [who uses it]. The main screen should show [X]. Users need to be able to [action 1] and [action 2]. Keep it simple — no login yet.
Mastery signal: You can explain what Lovable generated and point at which part of the screen came from which part of your prompt.
Day 2
Prompting for changes that stick
Learn the difference between a prompt that fixes one thing and one that rewrites your app.
Concepts
- · Be specific about the target: name the page, section or component you mean
- · One intent per message beats a five-part wishlist
- · Say what should stay the same, not just what should change
- · Use chat mode to plan a big change before letting it edit
Exercise
Make three deliberate changes: one layout change, one copy change, one behaviour change. Each in its own message. Note which phrasing produced a clean result.
Try this prompt
On the [page name] page, change only the [specific section]: [what you want]. Don't change the layout, styling or behaviour of anything else on the page.
Mastery signal: Your last three prompts each changed exactly what you intended and nothing else.
Day 3
Visual edits, design system and speed
Stop spending prompts on things you can click, and get your design consistent instead of ad-hoc.
Concepts
- · Visual edits let you change text, spacing and colours directly — free and instant
- · Ask for a design system (colours, fonts, tokens) once instead of restyling page by page
- · Reference images: paste a screenshot and describe what you want borrowed from it
- · Consistency comes from tokens, not from repeating style instructions in every prompt
Exercise
Fix five small visual issues with visual edits only, no prompts. Then use one prompt to define your palette and typography across the whole app.
Try this prompt
Define a design system for this app: a [mood] palette, [heading/body] typography, and consistent spacing and button styles. Apply it across every existing page — don't change any functionality.
Mastery signal: You reach for visual edits first and only prompt when the change is structural.
Day 4
Data, accounts and files
Turn a static prototype into something that stores real information.
Concepts
- · Enabling the backend gives you a database, authentication, file storage and server functions
- · Tables need access rules — who can read and write which rows
- · Auth first, then per-user data: it's far harder to retrofit
- · Never paste API keys into chat as plain text; use the secrets flow when prompted
Exercise
Add sign-up and login, then make one existing screen show only the signed-in user's own records. Create two test accounts and confirm they can't see each other's data.
Try this prompt
Add email and Google sign-in. Then make [feature] user-specific: each signed-in user should only see and edit their own [records]. Add the right access rules so no user can read another user's data.
Mastery signal: Two test accounts see completely separate data, and you verified it yourself.
Day 5
Debugging without burning credits
Read the error, isolate the cause, and fix it in one prompt instead of five.
Concepts
- · Paste the actual error text — screenshots of red boxes are weaker input
- · Say what you expected, what happened, and the exact steps to reproduce
- · If two attempts fail, switch to chat mode and ask for a diagnosis before another edit
- · Use version history to roll back rather than prompting your way out of a broken state
Exercise
Deliberately break something small, then fix it using one well-formed bug report. Then practise rolling back to a previous version.
Try this prompt
When I [exact steps], I expect [expected] but I get [actual]. Here's the error: [paste]. Diagnose the cause before changing anything, then fix only that.
Mastery signal: You've recovered from a broken state twice without starting the project over.
Day 6
Publishing and connecting the real world
Get the app in front of actual users on a real address.
Concepts
- · Publishing deploys the current version; preview and published are separate
- · Custom domains on paid plans; publish again after each change you want live
- · Connectors and integrations for email, payments, analytics and third-party data
- · SEO basics — page titles, descriptions, sitemap — matter the moment it's public
Exercise
Publish, open the live URL on your phone, and give it to one person who has never seen it. Write down every place they hesitated.
Try this prompt
Add proper page titles and meta descriptions for every page, a sitemap, and a robots.txt. Then tell me what else is missing before this is ready to share publicly.
Mastery signal: A stranger completed your app's main task without you explaining anything.
Day 7
Credits, limits and knowing when to leave chat
Build sustainably: understand what costs what, and where Lovable stops being the right tool.
Concepts
- · Credits are consumed per message, not per minute — planning in chat mode is cheaper than trial and error
- · Batch related small changes; visual edits are free
- · You can connect the codebase to GitHub and edit in a normal editor when you need precision
- · Honest limits: very complex logic, heavy data processing and niche integrations still need a developer's eye
Exercise
Review your last week of messages. Mark which ones were wasted, and write two rules for yourself to avoid repeating them.
Try this prompt
Review this project and list what's fragile, missing or likely to break as usage grows — prioritised, with the smallest fix for each.
Mastery signal: You can estimate roughly how many messages a change will take before you send the first one.