Introducing subagents: Lovable is now better at multitasking

Lovable has introduced subagents to enhance the efficiency and speed of its operations, particularly in larger projects. These subagents function like a kitchen team assisting a head chef, enabling parallel processing of tasks such as researching parts of a codebase or exploring the web, without altering the code itself. By distributing the workload, Lovable can provide faster builds and more precise answers while maintaining the integrity of the project's code. This read-only approach allows subagents to freely explore and report back to the main agent, which then makes informed decisions. Users don’t need to alter their interaction with Lovable, as subagents are automatically deployed and can be monitored in real time. This advancement not only reduces costs by routing lighter tasks to cheaper models but also ensures that Lovable performs more efficiently on complex tasks, offering deeper insights into larger codebases.