
The most driven founders are problem solvers. Watch their unscripted conversations with the Anthropic engineers.
Lovable turns a plain-language description of an app into production-grade software. Users shape it through a back-and-forth conversation, refining until it matches what they had in mind. Since its launch in November 2024, people have built more than 50 million projects on Lovable.

The most driven founders are problem solvers. Watch their unscripted conversations with the Anthropic engineers.
The most driven founders are problem solvers. Watch their unscripted conversations with the Anthropic engineers.
The most driven founders are problem solvers. Watch their unscripted conversations with the Anthropic engineers.

Learn when to use Haiku, Sonnet, or Opus to get better results and stay inside your rate limit. A practical guide to picking the right Claude model.
Learn when to use Haiku, Sonnet, or Opus to get better results and stay inside your rate limit. A practical guide to picking the right Claude model.
Learn when to use Haiku, Sonnet, or Opus to get better results and stay inside your rate limit. A practical guide to picking the right Claude model.

Anton is precise about what Lovable is. "We're not in the business of code generation," he says. "We're in the business of unlocking new economies and empowering every human out there who wants to create something. We take something that's super complex, and we try to make it seem very, very simple."
That planning capability is core to the experience. Lovable's first version took instructions and carried them out. Then came chat-based planning, where people work through what to build and how to structure it before any code is written. As that layer matured, the product began to feel like a partner. “We start seeing Lovable more as a co-founder than an AI software engineer,” Osika said. It decides how to act on the user's behalf, sometimes working quickly and sometimes taking longer, then telling the user it's done and asking what to change.
Alexandre Pesant, who leads product at Lovable, has watched that capability deepen with each model. “We've followed each model update because each model is better than the previous one,” he said. “Claude has a strong combination of great coding abilities and conversational abilities.”
Under the surface, Lovable runs on an agentic architecture. “We have a harness around a main agent that can use subagents to orchestrate tasks effectively, with the right models at each step,” said Pesant. The main agent reasons about a build and hands smaller pieces of work to subagents, matching each task to the model best suited to it.
Specific Claude releases marked turning points in what users could build. “Claude Sonnet 3.5 was the first model that made agents work,” said Pesant. “Claude Opus 4.5 was the next big step change in reliability on long-horizon tasks, unlocking a new class of projects.”
Every new Claude release goes through the same evaluation Lovable has run from the start, measuring how often the system hits a wall and produces an app that’s broken or isn’t what the user asked for. That gate matters because Lovable’s users often can’t read the code themselves, so they’re trusting the output to work. For Osika, earning that trust is the harder, more durable goal. “Something that I think is very, very rare in AI is a trusted brand that people love and keep coming back to,” he said. “To have that trusted brand, that's not something that you magically achieve.”