Lovable Ships 15 Updates Including MCP Integration and AI Image Generation.

Lovable announced over 15 product updates in the past three weeks, headlined by MCP server connectivity that pulls requirements, tickets, and wireframes directly from tools like Jira, Confluence, Notion, Linear, and n8n. The updates also introduce a dedicated Design view, theme management, AI-powered image generation, and enhanced privacy defaults for workspace projects.

The chat-based development platform, which lets users build full-stack web apps through conversation, detailed the changes in a LinkedIn post and its official changelog.

The Model Context Protocol (MCP) integration connects Lovable to work tools through n8n workflows, which support 400+ applications including Salesforce, Slack, and Google Sheets. Users can ingest product requirement documents, tickets, and design files without manually copying content into the chat interface.

This follows Lovable's pattern of expanding connector support, allowing teams to build from existing documentation and project management systems. The MCP implementation means requirements stored in Linear or Confluence can flow directly into the build process.

Lovable introduced a dedicated Design view that consolidates visual editing, themes, and image generation in one panel. The themes feature lets users set brand standards for colors, typography, and spacing once, then apply them across any project with a preview option before committing changes.

The visual editing tools now support multi-element selection and mixed-content editing. Users can modify text within buttons that contain both text and icons, adjust margins, padding, borders, shadows, and colors simultaneously across selected elements. The interface handles icons and AI-generated images alongside traditional text editing.

The platform added AI-powered image generation that creates custom visuals based on text descriptions. Users can generate original images directly in the editor without consuming Lovable credits, the platform's usage currency. This removes a friction point for teams prototyping visual-heavy applications or needing placeholder graphics during development.

All Lovable plans now default to workspace-only project visibility. Projects remain private to the workspace unless users explicitly change visibility settings or publish the application. This addresses a common concern among teams building internal tools or client work where public exposure isn't desired.

The Shopify integration received permission scoping updates. Only the user who connected the Shopify store maintains write access, while collaborators get read-only access to store data but retain full ability to build storefront functionality. This prevents accidental modifications to production stores during collaborative development.

Lovable now supports Google's Gemini 3 Pro and Nano Banana Pro models alongside existing Claude and GPT options. The platform upgraded to Claude Opus 4.5 earlier this year for improved design output, and adding Gemini gives users more model choices for different build scenarios.

The full list of updates includes improvements to error handling, performance optimizations, and UI refinements detailed in the platform's documentation. Lovable continues competing with tools like Bolt, v0, and Replit in the AI-assisted development space, differentiating through its chat-first interface and integrated deployment pipeline.