Model Context Protocol
Build and analyze onboardingfrom your AI assistant
Produktly's MCP server gives Claude, ChatGPT, Cursor and other AI tools read and write access. They can create and update product tours, checklists, smart tips, announcements, NPS widgets, micro-surveys, feedback widgets and changelog posts, and read your analytics, feedback and roadmaps. New items are drafts by default and go live only when published with targeting.
What can you do with it?
Give your AI assistant direct access to your product data and unlock new workflows
Build tours, checklists and surveys
Ask your coding agent to add a product tour to the page it just built, a setup checklist, an NPS survey or a smart tip. It creates them in Produktly with steps, CSS selectors and targeting, as drafts by default, and can publish them or update them later.
Analyze widget performance and feedback
Ask your AI which tours have the best completion rate, how engagement changed this week vs last, what detractors are saying, or to summarize recurring feedback themes. Turn usage data and customer voices into actionable insights in one conversation.
Draft and publish changelog posts
Let your AI write changelog posts directly in Produktly from your git history, release notes, or a quick description. Leave them as drafts for you to review, or publish them straight away when you already know what shipped.
Prioritize your roadmap
Ask your AI to cross-reference roadmap items with feedback and vote counts to help you decide what to build next. Make data-driven prioritization decisions.
Prepare for meetings
Get a quick summary of recent feedback, shipped features, and roadmap progress before your product review. Your AI has the full picture in one conversation.
Answer questions instantly
"When did we ship feature X?" "What feedback did we get on the new dashboard?" "How many votes does the dark mode request have?" Get instant answers.
Connect the dots
Your AI can pull data across changelogs, feedback, roadmaps, and NPS in a single conversation. Understand the full story from user request to shipped feature.
What your AI assistant can do
Create and update (write)
Product tours, checklists, smart tips, announcements, NPS widgets, micro-surveys, feedback widgets and changelog posts. Items are created as drafts unless you ask to publish, and publishing needs targeting. Updates overwrite the fields you change.
List and read
Every widget type, widget stats and a summary across widgets, NPS scores and responses, feedback responses, roadmaps, tags, and a setup guide for installing the script.
Access follows your role: workspace viewers connect read-only, editors get read and write. With a private key, the key's scopes decide.
Get started in minutes
Connect your favorite AI assistant to Produktly in three simple steps
1. Add the server URL
2. Sign in with Produktly
3. Ask it to build or analyze
Simple configuration
Claude, ChatGPT and Gemini CLI only need the URL and a Produktly login. Other MCP clients can use a private key:
MCP client configuration with a private key
{
"mcpServers": {
"produktly": {
"type": "streamable-http",
"url": "https://api.produktly.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_PRIVATE_KEY"
}
}
}
}Works with your favorite AI tools
Any AI assistant that supports the Model Context Protocol can connect to Produktly