Lovable's MCP Server Lets Claude Build and Ship Apps Without Human Handoffs
Lovable's new MCP server lets Claude chat, Claude Code, and Claude Design all drive your Lovable workspace directly — collapsing the gap between ideation and shipped app.
- Lovable launches a native MCP server at lovable.dev/mcp, letting Claude chat, Claude Code, and Claude Design all drive your Lovable workspace programmatically.
- Claude chat users can add Lovable as a connector via Settings → Connectors → Add custom connector →
https://mcp.lovable.dev. - Claude Code users get
/build,/deploy, and/dbcommands, enabling full Lovable project management from the terminal. - Claude Design export now includes Lovable as a destination, letting you go from a visual prototype to a deployed full-stack app with one click.
- Available on all Lovable plans including Free; only
create_projectandsend_messagecalls consume Lovable credits — all other tools are free. - OAuth only — no API keys; currently supports Claude (chat/desktop), Claude Code, Cursor, and VS Code as clients.
The wall between your AI chat window and your app-building platform just got a lot thinner. Lovable has launched a native MCP server that lets Claude (the chat assistant), Claude Code (the terminal agent), and Claude Design (the visual prototyping tool) all connect directly to your Lovable workspace. You can now scaffold a project, iterate on it, query its live database, and ship it to a public URL without ever opening the Lovable UI.
Three integrations, one announcement
The release bundles three distinct integration points, each targeting a different part of the development workflow:
- Claude chat connector: Add Lovable as a connector inside claude.ai or Claude Desktop. From there, you can audit projects, query live databases, and start new builds entirely through conversation.
- Claude Code terminal commands: Run
claude mcp add --transport http lovable "https://mcp.lovable.dev"from your terminal and Lovable becomes a first-class tool in your agentic coding sessions. The integration ships/build,/deploy, and/dbas single commands. - Claude Design export: Claude Design , Anthropic's visual prototyping tool that launched in April 2026 , can now export directly to Lovable, letting you go from a static mockup to a full-stack app with auth, a database, and a live URL in one click.
What MCP actually means here
MCP (Model Context Protocol) is an open standard that lets AI agents discover and call external tools. When an agent connects to an MCP server, it can see what tools are available and decide when to use them. Lovable's MCP server makes Lovable one of those tools , findable, invokable, and composable. Think of it as giving your AI client a remote control for your entire Lovable workspace.
The key distinction worth understanding: chat connectors let the Lovable agent connect to your external tools (Notion, Linear, Miro, and so on) during a build. The Lovable MCP server is the reverse , it lets your assistant or editor connect to Lovable and drive it programmatically. The direction of control is flipped.
What the MCP server can actually do
The Lovable MCP server exposes the same primitives Lovable runs on: projects, chat, files, connectors, databases, and deploys , as tools any agent can call. In practice, that covers a lot of ground:
- Create projects from a prompt: Spin up a new Lovable app from your agent. One tool call, full project scaffolded with React, Tailwind, and shadcn/ui.
- Iterate via chat: Tell Lovable what you want changed. It reads the code, makes the edits, and rebuilds the preview.
- Inspect code: List files, read source, and review the diff from every agent response.
- Manage connectors: Browse Lovable's connector catalog and wire up Gmail, Notion, Linear, and more from inside your agent.
- Database operations: Your workspace's data plane, configured by chat. Enable Lovable Cloud, run live SQL, and pull connection info , all from your agent. Reads, writes, and schema changes.
- Deploy: Trigger a deploy from your agent and get back a live URL. Iterate and share, all programmatically.
- Analytics: Pull real-time visitor counts and historical pageview data for any deployed project.
Getting connected
Setup varies by client. For Claude chat,
go to Settings → Connectors, click Add custom connector, set the name to Lovable and the URL to https://mcp.lovable.dev, click Add to save, then sign in to Lovable when Claude prompts you to authenticate. The Lovable tools appear in the chat composer's tool menu once authentication is complete.
For Claude Code users, a single terminal command does it:
claude mcp add --transport http lovable "https://mcp.lovable.dev"
OAuth is the only supported way to connect to the Lovable MCP server.
No API keys to manage.
Using the MCP server does consume your Lovable credits for create_project and send_message calls. Other tools are free.
The Claude Design pipeline
The third integration is arguably the most interesting for product teams. Claude Design now sends work to Adobe, Base44, Canva, Gamma, Lovable, Miro, Replit, Vercel, and Wix, in addition to PDF and PowerPoint. The breadth of this list reveals a deliberate positioning strategy: Anthropic is building Claude Design not as a place where work is finished, but as the place where it begins.
Claude Design is Anthropic's visual creation tool that turns prompts into prototypes, slides, one-pagers, and landing pages. It launched in April 2026 on Pro, Max, Team, and Enterprise plans. The Lovable export path means a designer can sketch a UI in Claude Design and hand it off to Lovable to add a real backend , auth, database, payments , without any manual code transfer.
Who this is actually for
The MCP server is aimed at developers using ChatGPT, Claude Code, Cursor, or VS Code who want to spin up and iterate on Lovable projects without leaving their AI client or editor, and at teams who want Lovable as a step in a larger agent flow: generate content, scaffold a UI, deploy it, and hand off the URL.
The Claude Code path in particular closes a gap that the developer community had been navigating manually. The dominant workflow before this announcement was: build a prototype in Lovable, export to GitHub, then pick it up in Claude Code for production-grade work. The pattern that works is Lovable for initial scaffolding and UI, then migrate to GitHub and Claude Code for everything after. Multiple Reddit users describe this as the de facto hybrid workflow. The MCP integration makes that handoff programmatic rather than manual.
Limitations worth knowing
A few constraints to keep in mind before wiring this into a production pipeline:
- VS Code and other MCP clients cannot complete the OAuth flow at this time. Only Claude (chat and desktop), Claude Code, and Cursor are fully supported.
- The MCP server inherits your Lovable user permissions exactly, with no additional surface area , but that also means a connected agent has access to every project in your workspace, not just one.
- The Lovable MCP server is available on every plan, including Free. Enterprise customers should reach out to their account executive to enable it for their workspace.
- The known production limitations of Lovable itself still apply. The Superblocks team concluded: "Lovable gets you at most 70% of the way there, but you'll spend a lot of time wrestling with that last 30% to make it usable for real customers. Security and data handling still feel immature, and debugging inside the platform can trap you in frustrating prompting loops."
The bigger picture
Lovable hit $400 million in ARR in just over a year , and this integration signals that the company is betting its next phase of growth on becoming infrastructure for the agentic development ecosystem rather than just a standalone app builder. By positioning itself as a tool that any MCP-compatible agent can call, Lovable becomes composable , a deployment and scaffolding primitive that fits inside larger automated workflows.
Claude Design's transformation from prototype toy to enterprise platform is the latest move in a company-wide strategy to make Claude not just an assistant people talk to, but a worker embedded in the systems where work actually happens. The Lovable integration is one concrete expression of that strategy: the idea-to-shipped-app pipeline now has fewer manual handoffs than it ever has before.