cloudparts

LovableRenamed

A Stockholm-built prompt-to-app service that generates a TanStack Start web app with a built-in Supabase-based backend, hosts it, and syncs the code to your Git repository.

What it is
AI app builders
Type
Commercial product
Made by
Lovable Labs Incorporated
License
Proprietary
Status
Renamed. First known as GPT Engineer, after Anton Osika’s open-source gpt-engineer CLI; it launched as Lovable in November 2024.
Facts checked
13 October 2026

What it is

You describe an app and Lovable builds the whole thing: frontend, database, authentication, storage and server functions. Every project uses the same stack, with no choice of framework or database. Apps created since 13 May 2026 use TanStack Start with server rendering; older ones use React and Vite. Data lives in PostgreSQL, either in Lovable Cloud, a managed backend built on Supabase’s open-source foundation, or in a Supabase project you connect yourself. Cloud adds a database with security policies you can review, sign-in (email, phone, Google, Apple, Microsoft, SAML), file buckets, edge functions, scheduled jobs and logs, hosted in the Americas, Europe or Asia Pacific; the region can’t change once chosen.

The project chat has three modes. Chat mode discusses changes without making them, Plan mode writes a plan you approve, and Build mode (formerly Agent mode) explores the codebase, edits across files, reads logs and network activity, and verifies its work. By default it may change your database schema and data without asking, but changes that could break the published app, such as dropping or renaming a column, are blocked unless you allow them. Lovable picks the model itself; you can’t choose one for building.

Usage is paid in credits. A message costs according to the work done: Lovable’s examples range from 0.50 credits for “Make the button gray” to 1.70 for a landing page with generated images, and Plan mode costs 1 credit per message. Every plan gets 5 free build credits a day (capped at 30 a month on Free) and monthly grants for hosting and AI features. Code can be edited and downloaded on paid plans, and two-way Git sync with GitHub, GitLab or Bitbucket is available on every plan.

Why teams pick it

  • A full backend with no setup: database, auth, storage, functions and jobs appear when a prompt needs them
  • Your code is portable: two-way Git sync keeps it in your own repository, and you can deploy it elsewhere
  • Guardrails around the database: backward-incompatible migrations are blocked by default, and security scans check policies and dependencies
  • Enterprise controls: SSO, SCIM, SOC 2 Type II, ISO 27001, data region pinning and EU inference

Watch out for

  • One fixed stack: web apps only (no native mobile), TanStack Start or older React and Vite, and PostgreSQL through Cloud or Supabase
  • You can’t pick the model behind Build, Plan or Chat mode
  • Since 9 September 2026, Free and Pro workspace data (prompts, code, outputs) may be used to train Lovable’s models unless you opt out; Business and Enterprise are excluded by default
  • A Cloud project’s region is fixed when it is created and can’t be moved later

Alternatives

What all ai app builders have in common →Compare side by side →

How it fits with other technologies

See it in a real stack

No architecture uses Lovable yet. Drop it into the playground and wire it up yourself.Open playground →
Where it came from2023–2026

In April 2023 Anton Osika published gpt-engineer, an open-source command-line tool: specify software in natural language, then watch an AI write and run the code and ask it for improvements. The repository now has more than 55,000 stars on GitHub. Osika and Fabian Hedin founded a company around the idea in Stockholm in 2023, first called GPT Engineer, which built a managed service at gptengineer.app; the CLI’s README pointed to that service as its evolution.

The service launched as Lovable in November 2024. Lovable added its own hosted backend, Lovable Cloud, an agentic Build mode, Git sync and enterprise controls. It raised a Series B in December 2025 and, on 12 August 2026, a $400 million Series C at a $13.3 billion valuation led by Menlo Ventures. By then people had created more than 60 million projects, Lovable-built apps drew more than 900 million visits a month, and Lovable said employees at nearly two-thirds of the Fortune 500 used it.

In 2026 the stack changed: new apps from 13 May use TanStack Start with server rendering instead of a browser-rendered React and Vite app. The gpt-engineer repository is now archived and describes itself as the precursor to Lovable.

Why it’s built this way

A managed service instead of a CLI
The gpt-engineer README split the audience: people who wanted an opinionated, managed service were sent to gptengineer.app, and people who wanted a hackable CLI to aider. The company built the managed service.
One stack for every project
Lovable doesn’t offer a choice of framework or database, so its agent, backend and hosting all target the same stack; anyone who needs another one syncs the code to Git and continues outside Lovable.
Supabase’s open-source foundation for the backend
Building Cloud on Supabase gives every app PostgreSQL, auth, storage and realtime from day one without users setting up a separate Supabase project, while the same projects can still connect to Supabase directly.

Timeline

  1. 2023

    Anton Osika publishes the gpt-engineer CLI; the company starts in Stockholm as GPT Engineer

  2. 2024

    Launches as Lovable in November

  3. 2025

    Series B in December

  4. 2026

    New apps move to TanStack Start on 13 May

  5. 2026

    $400M Series C at a $13.3B valuation on 12 August

Sources10 · facts checked 13 October 2026

Improve this page on GitHub