This guide demonstrates how to use the Unbounce Model Context Protocol (MCP) integration to create landing pages, analyze campaign performance, and manage your account using plain-language AI prompts.
Prompt your Pages to Life
Build a landing page from scratch
Just ask: “Use my brand guidelines to build a landing page for our summer sale with an email signup.”
What happens: Your AI assistant reads your brand guidelines, builds the page in your branding, and stages it as a draft. Confirm, and it publishes live with the form ready for lead capture.
Build a landing page from a Figma design
Just ask: “Build a lead gen landing page using this Figma design as a template.”
What happens: Your AI assistant reads the design file and rebuilds it as a live page—layout, type, color, and copy intact. Your designer's work ships instead of waiting in a dev queue. (Requires the Figma MCP server. Image download/upload requires Claude Code.)
Build a landing page from a brief
Just ask: “Use this document to build a landing page for our Black Friday campaign.”
What happens: Your AI assistant reads the brief—conversion goal, audience, key messages—and builds the page against it. The campaign brief becomes the page.
Build and style forms on MCP-created pages
Just ask: “Add a multi-step form to the summer sale page. Include first name, last name, phone, business email, and company size (optional).”
What happens: Your AI assistant writes the form directly into the page—fields, labels, validation, and multi-step logic—and Unbounce wires it for lead capture automatically.
Add interactive elements on MCP-created pages
Just ask: “Add a testimonial carousel to the social proof landing page.”
What happens: Pages built this way are custom HTML, CSS, and JavaScript under the hood, so your AI assistant can add carousels, accordions, countdown timers, or sticky CTAs without a plugin or a developer. Describe the change and it rewrites the element in place.
Put landing page insights within reach
Compare conversion rates across pages
Just ask: “Compare conversion rates across every landing page in the Summer Sale group and tell me which ones are underperforming.”
What happens: Your AI assistant pulls visits, visitors, conversions, and conversion rate for every variant—plus the lift and statistical confidence between them—for any date range you name, and identifies which is underperforming.
Add a variant and split traffic
Just ask: “Add a second variant with a shorter headline, and split traffic 50/50.”
What happens: Your AI assistant creates the challenger variant and sets the traffic split.
Read your leads and stats
Just ask: “Show me the leads and conversion rate for the webinar page.”
What happens: The AI assistant returns each captured lead (with the variant it came from) and the page’s visit/conversion stats.
Make landing page management effortless
Update elements and copy on MCP-created pages
Just ask: “Change the buttons to blue and update the footer copyright to 2026 on the Forgeloop page.”
What happens: Your AI assistant makes the edit and restages the page as a draft. Review it, then publish when you're ready.
Organize pages into groups
Just ask: “Create a group called Summer Campaign and put these three pages in it.”
What happens: The AI assistant creates the page group and adds the pages. Groups are pure organization and take effect immediately.
Unpublish or delete a page
Just ask: “Unpublish the old promo page,” or “Delete the duplicate draft.”
What happens: Unpublishing takes a page offline (reversible by republishing). Deleting is a soft-delete to trash, restorable from the Unbounce app.
Update metadata across pages
Just ask: “What published pages are missing metadata? Suggest copy, and update.”
What happens: The AI assistant scans your published pages for missing metadata, then suggests copy based on the landing page content and metadata best practices.