Skip to content
Get packs

Open Lovable

Chat with an AI to build and recreate React apps, with a Firecrawl-made example app.

GitHub Download source
Stars
28.8k
Forks
5.5k
Licence
MIT
Language
TypeScript
Last update
28 days ago
Type
App

What it is

Open Lovable is a web app where you chat with an AI model to build React apps. It is described as an example app from the Firecrawl team, and the project description says it can clone and recreate a website as a modern React app. The README is short and does not fully explain how the cloning works beyond the setup steps.

  • Run the app locally and chat with an AI model to build a React app
  • Point it at a website to attempt a React recreation of that site
  • Switch between supported AI providers (Gemini, Anthropic, OpenAI, Groq) by setting their keys
  • Use E2B instead of the default Vercel Sandbox by changing the sandbox provider setting

Great for

  • Developers who want a self-hosted example of an AI app builder
  • Developers exploring how Firecrawl and an LLM can be combined
  • People who already have API keys for one of the supported AI providers
  • Designers who want a quick React starting point from a website they like

Not ideal for

  • Non-developers who want a hosted, no-setup builder: the README points to Lovable.dev for a complete cloud solution
  • People without API keys or accounts for Firecrawl, an AI provider, and a sandbox provider, since setup needs several of these
  • Teams needing production-ready code: the README does not describe production readiness
Step-by-step guide: how to use Open LovableInstall, first run, examples, tips and FAQs
Open Lovable vs OpenHandsBolt.new vs Open Lovable

How to get it

  1. 1

    Clone and build from source

    You'll need: Node.js · pnpm, npm, or yarn · Git

    Clone the repository

    git clone https://github.com/firecrawl/open-lovable.git

    Enter the project folder

    cd open-lovable

    Install dependencies

    pnpm install
  2. 2

    Run it

    • Clone the repository and install dependencies
    • Create a .env.local file and add your Firecrawl API key and at least one AI provider key
    • Choose a sandbox provider: Vercel (default) or E2B, and add its credentials
    • Run pnpm dev
    • Open http://localhost:3000 in your browser
  3. 3

    Get the most out of it

    • Add several AI provider keys to choose between models in the app
    • Set MORPH_API_KEY for the optional Fast Apply feature to speed up edits
    • Use vercel link and vercel env pull to get the Vercel OIDC token automatically
    • Switch SANDBOX_PROVIDER to e2b if you prefer E2B over Vercel Sandbox

Downloads

Source code (.zip)main branch · straight from GitHub

Strengths

  • MIT licensed, so the code can be reused and modified
  • Supports several AI providers through environment variables
  • Clear setup steps with an example environment file in the README

Trade-offs

  • Requires multiple third-party API keys and accounts to run fully
  • The README is brief and does not document architecture, limits, or costs
  • Vercel sandbox setup has two authentication methods, which adds configuration decisions

How it compares

ProjectStarsTypeLevelUpdatedBest for
open-lovable (this)28.8kAppintermediate28 days agoDevelopers who want a self-hosted example of an AI app builder
bolt.new16.6kAppbeginner1.9 years agoPeople who want to prototype a full-stack web app quickly without local setup
OpenHands92.5kAppintermediate4 days agoDevelopers who want coding agents running on a server while their laptop is off

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.