Skip to content
Type
prompt library
Verdict
Useful
Licence
Conditional licence, mixedmixedThe public prompts are free; the bundle is $199 once and includes a commercial-use licence, Pro channel versions and four build kits whose row-level security was executed and reviewed
Agent channels
prompts
Pricing
freemium
Reviewed

← Atlas · Topics: agents-and-prompts, landing-pages

Lovable Prompts ​

Type
prompt library
Verdict
Useful
Licence
Conditional licence, mixedmixedThe public prompts are free; the bundle is $199 once and includes a commercial-use licence, Pro channel versions and four build kits whose row-level security was executed and reviewed
Agent channels
prompts
Pricing
freemium
Reviewed

What it is ​

A prompt library with one discipline most of its neighbours lack: it publishes what a prompt cost to run. Four builds are measured in Lovable's own credits, read from the balance before and after, with dates — a landing page and email opt-in at 1.8 credits, empty states and skeletons at 9.5, a SaaS MVP with auth at 10.4, a built-in SEO review at 1.1.

The prompts themselves are structured rather than conversational: context, numbered features, the screen states a builder skips unless told (empty, loading, error), a build order, safeguards, and a closing instruction that makes the tool ask clarifying questions before writing code. The Pro versions add a growth channel — organic search, product-led invites, outreach — so the build arrives with a distribution assumption instead of a launch project left for later.

The library is 122 free prompts across eight categories (app builder, UI/UX, backend logic, SaaS, landing page, API integration, CRM, admin dashboard) with 152 in the paid bundle, and a free generator that gives three prompts to a free account.

When to open it ​

Open it when you are about to paste a one-line brief into an AI builder and want a prompt that names states and questions instead of describing a vibe, or when you want to compare what things cost to generate before spending a plan's credits.

Most useful ​

  • Measured credit costs with dates, which is the most honest number in this whole category.
  • States written into the prompt — empty, loading, error — because those are exactly what a generator omits.
  • A closing line that forces questions first, a small trick that prevents a large rework.
  • Build order inside the prompt, so the agent does not start with the least important screen.
  • A free generator with a quota, rather than a wall.
  • Guides with dated prices for seven builders, which age better than comparisons.

Using it with agents ​

Prompts are the channel: the page publishes them in full with copy buttons, so an agent can fetch the text of a public prompt and adapt it. There is no llms.txt at review and no API. The bundle's Pro versions are behind a payment, and the generator requires an account after three.

Watch out for ​

  • The site earns commission on Lovable sign-ups and says so; the "try it" links are affiliate links.
  • Four measured runs is a sample, not a benchmark, and credits vary with the plan and the app.
  • Growth channels written into a prompt are opinionated (product-led invites, outreach) and may not fit the product.
  • The bundle price is one payment but the prompts are tied to one builder's prompting style.
  • Generated apps still need review: a prompt that names states does not guarantee the states work.

Reusable ideas ​

  • Publish what a prompt cost to run, with a date, as part of the prompt.
  • Name the states a generator forgets, in the prompt, in order.
  • End every build prompt by asking the tool what it needs to know.
  • Put the distribution plan in the same prompt as the product.

Jiro, OnePromptDesign, VibePrompts