The MCP built for
agentic development
Build, iterate, and deploy full-stack apps on Lovable from the AI client you already work in. Powered by the Model Context Protocol.
Use Lovable to build me a SaaS landing page for an AI receipt scanner. Add a hero, pricing, and a waitlist form wired to a Supabase table.
- Calling create_project…
- Calling send_message— “Build the landing page with hero, pricing, waitlist…”
- Calling read_file on src/pages/Index.tsx
- Calling execute_sql — creating waitlist table
Preview ready at receipt-scanner.lovable.app
Works with AI assistants, Cursor, and VS Code · Available on every plan.
Everything an agent needs to ship a real app
The Lovable MCP server exposes the same primitives Lovable runs on: projects, chat, files, connectors, databases, and deploys, as tools any agent can call.
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.
Read files, inspect diffs
List files, read source, and review the diff from every agent response.
Add Lovable connectors
Browse Lovable's connector catalog and wire up Gmail, Notion, Linear, and more from inside your agent. Your workspace's data plane, configured by chat.
Lovable Cloud database
Enable Lovable Cloud, run live SQL, and pull connection info, all from your agent. Reads, writes, and schema changes.
Deploy to production
Trigger a deploy from your agent and get back a live URL. Iterate and share, all programmatically.
Workspaces & projects
List projects, fetch metadata, manage workspaces. Everything an agent needs to operate across all your Lovable projects.
Persistent knowledge
Set workspace and project knowledge so Lovable follows your standards, brand, and API patterns on every run.
Built for agents
Typed tool schemas, structured outputs, and clear error responses. Designed for agents from day one.
Connect your client
Pick your client and follow the steps. Authentication uses OAuth, so there are no API keys to manage.
Add Lovable through Claude's connector settings. Works in Claude Desktop (macOS and Windows) and on claude.ai.
- 1In Claude, go to Settings → Connectors.
- 2Click Add custom connector.
- 3Set the name to
Lovableand the URL tohttps://mcp.lovable.dev. - 4Click Add, then sign in to Lovable when Claude prompts you.
The Lovable tools appear in the composer's tool menu once authentication is complete.
Frequently asked questions
What is MCP?
The 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.
Which clients can I use?
OAuth is currently supported for the assistant and editor clients listed in the setup section below. Other MCP clients cannot complete the OAuth flow at this time.
What's the difference between this and Lovable's chat connectors?
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.
Is it included in my plan?
Yes — 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.
Can I connect with an API key?
Not currently. OAuth is the only supported way to connect. The MCP server uses your Lovable user permissions — it can only access workspaces you already have access to.
