Lovable built its reputation as an AI-powered platform for creating applications. Yet the company is increasingly positioning itself for a future in which fewer people rely on traditional user-facing apps — driven by the rise of agents.
In a recent blog post, Lovable described a vision of “a digital brain for your team connecting your daily tools.” Fabian Hedin, Lovable’s chief technology officer, told Latent Space in an interview: “you can get to a place where you’re using one entry point to all the work that you’re doing.”
What is a "capability"?
Lovable still wants developers to use its platform to build applications, but it is putting growing emphasis on what it calls “capabilities.” The company defines a capability as a useful part of an application that an agent can call directly, without a human having to open the full app.
Lovable can convert a published application into agent-accessible capabilities by exposing selected functions as tools through a hosted MCP (Model Control Protocol) server. The outcome is effectively a single application with two interfaces: a conventional human UI and a new agent interface that MCP-compatible AI clients such as ChatGPT and Claude can use.
Rapid product evolution in three years
Lovable’s trajectory over the past three years has been fast. The company grew out of GPT Engineer, an open-source coding tool launched in 2023 that initially focused on prototyping. In November 2024 it became a commercial product, and in December 2024 it was rebranded as Lovable.
The team noticed early on that some users were building production-grade apps on the platform — not just prototypes or MVPs but actual products serving customers. Fabian Hedin said they also saw users creating internal software, sometimes to support a public-facing product and other times as enterprise internal tools such as CRM systems, admin panels, or support consoles.
As a result, Lovable has evolved into an end-to-end software creation and hosting company, operating in some of the same spaces as firms like Vercel and Cloudflare, although Lovable emphasizes AI-generated software rather than infrastructure alone.
User and revenue growth
This fast product evolution has been accompanied by notable user and revenue metrics. Deedy Das, a partner at lead investor Menlo Ventures, tweeted that the company surpassed a $500 million annualized revenue run rate, has more than 60 million projects created, and sees over 900 million monthly visits to Lovable-built apps. Lovable also reports that employees at nearly two-thirds of the Fortune 500 have used the platform.
Menlo Ventures led Lovable’s $400 million Series C this month, with the Scaleup Europe Fund managed by EQT also participating, valuing the company at $13.3 billion.
Hedin credits the pace of change both to Lovable’s product innovation and the rapid improvements in large language models (LLMs): “Every few months, we introduce new capabilities at the application layer, while the large language models also continue improving. Those two things compound.”
The company "brain" concept
For Lovable, a digital brain means a single interface through which many different tools and workflows are accessible. Hedin says such a system should hold as much context as possible about the user, the company and the surrounding world, and it must include capabilities to perform both general tasks and organization-specific actions.
Ultimately, he added, the goal is that “everything that you’re building can be reused in an agentic way.” In practice, applications become collections of capabilities that users access increasingly via an organizational agent rather than—or in addition to—the app’s own UI.
“Our job as a platform is to ensure that all these separate capabilities are connected through one agent — not that you have to build a different agent for every task,” Hedin said. He gave an internal example: a Lovable-built tool for granting credits to users and managing platform functions now exposes those capabilities internally through the Lovable agent.
Lovable’s agent is also designed to work asynchronously: it can schedule itself to resume tasks later (for instance to check a deployment or monitor a recurring process) and then return the result to the same conversation.
Competition and differentiation
Lovable is not alone in pursuing a company-brain vision. Guillermo Rauch, CEO of Vercel, recently introduced Vercel’s internal agent, @𝚟, and tweeted that it is becoming part of daily workflows with exponential growth in interactions and token use.
Hedin acknowledged that Vercel and other AI companies are heading in similar directions, but argued Lovable’s wedge is “being the best place to build the capabilities that agents need.” In his view, orchestrating capabilities is the easy part; ensuring they are well connected, correctly built and reliable is the harder challenge.
He also explained why the team prefers the word “brain” to simply “agent”: the former emphasizes connecting appropriate context and capabilities rather than just modeling an agent as an employee-like task executor.
Security and connecting external systems
One of the biggest challenges for the agents-and-capabilities approach is security. If an employee builds an app on Lovable that connects to a company Slack, the platform must prevent accidental exposure of personal messages or other confidential data to the company brain.
Lovable uses connectors to link external tools and services. Hedin said the platform needs a “kind of permissioning graph” to maintain security and privacy. In a technical blog post, Lovable described an “app user connector” type that preserves each user’s identity and source-system permissions. Credentials are stored encrypted server-side and handled by Lovable’s connector gateway rather than being exposed to the generated application; the app receives a short-lived key bound to the relevant user instead.
“We separate the connection to external systems from the application code being written,” Hedin said. “The application interfaces with the Lovable platform, but the application itself never gets access to those credentials.”
What this means for the future of SaaS
Lovable is positioning for a future in which a company brain uses capabilities derived from the apps users build. That raises questions about the future of traditional SaaS apps. Hedin reiterated that people will increasingly interact with software through an AI layer — the company brain concept — which will consolidate many separate tools into a single experience while preserving the vertical capabilities those tools provide.
He accepts that some traditional SaaS vendors may resist by clinging to their existing app experiences, but Lovable’s aim is to be the platform for building capabilities and for connecting anyone who wants to use them. Hedin’s closing advice for SaaS businesses was pragmatic: they should focus on providing the “shovel” for AI to use their capabilities — in other words, make their functionality accessible and usable by agents.
