Lovable, the AI-powered app builder, just shipped a meaningful architectural upgrade: subagents. The main Lovable agent can now spin up a team of helper agents behind the scenes to research your codebase, look things up on the web, and review work in parallel , all without you changing anything about how you use the product.

The problem it's solving

On large projects, Lovable can be slow. You send a message and watch it churn before anything starts happening , a lot of that time is spent on discovery: reading through your codebase and understanding the code before making any changes. On larger projects, there's just more to look through: more files, more features, and more places a change could ripple to.

There's also a subtler problem: context window pollution. A context window is essentially an AI model's short-term memory , everything it's currently thinking about. The more you stuff into it, the harder it is for the model to stay focused. On long sessions with big codebases, this degrades output quality in ways that are hard to diagnose but easy to feel.

How subagents actually work

Each subagent gets a specific job , like looking through part of the project or researching something online , and comes back with what it found. Subagents can dig deep through your whole codebase or across the web. The one thing they can't do is change your code. That part stays with the main agent, by design: it means subagents can move fast and explore freely without any risk of touching your app.