MCP Server

Produktly provides a Model Context Protocol (MCP) server that lets you connect AI assistants like Claude, Cursor, Windsurf, and others directly to your Produktly data.

With the MCP server you can query your changelogs, feedback widgets, roadmaps, NPS widgets, and cross-feature usage stats through natural language — and have your AI assistant create and update changelog posts on your behalf.

Endpoint

https://api.produktly.com/api/mcp

The server uses the Streamable HTTP transport (stateless mode).

Authentication

The MCP server authenticates using your Produktly Private Key as a Bearer token.

1. Go to Settings > Private Keys in your Produktly dashboard
2. Generate a new private key (give it a name like "MCP")
3. Use the key as a Bearer token in your MCP client configuration

Setup

Claude Code (CLI)

Run this command in your terminal:

claude mcp add produktly --transport http --url https://api.produktly.com/api/mcp --header "Authorization: Bearer YOUR_PRIVATE_KEY"

Or add it manually to your Claude Code MCP settings:

{
  "mcpServers": {
    "produktly": {
      "type": "streamable-http",
      "url": "https://api.produktly.com/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_PRIVATE_KEY"
      }
    }
  }
}

Cursor / Windsurf

Add to your MCP configuration file:

{
  "mcpServers": {
    "produktly": {
      "url": "https://api.produktly.com/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_PRIVATE_KEY"
      }
    }
  }
}

GitHub Copilot / VS Code

In VS Code, open the Command Palette and run MCP: Add Server. Select HTTP as the transport type, then enter the URL https://api.produktly.com/api/mcp.

Or add it manually to your .vscode/mcp.json:

{
  "servers": {
    "produktly": {
      "type": "http",
      "url": "https://api.produktly.com/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_PRIVATE_KEY"
      }
    }
  }
}

The Produktly tools will be available in Copilot's Agent mode.

Other MCP clients

Any MCP client that supports the Streamable HTTP transport can connect. Set the URL to https://api.produktly.com/api/mcp and include the Authorization: Bearer YOUR_PRIVATE_KEY header.

ChatGPT / Gemini

ChatGPT and Gemini require OAuth-based authentication for MCP servers, which is not yet supported. We're working on adding OAuth support — stay tuned.

Available tools

list_changelogs

List all your changelogs (id, name, active status, creation date).

get_changelog_posts

Get posts for a specific changelog.

Parameter	Type	Required	Description
changelogId	number	Yes	The ID of the changelog
startDate	string	No	Filter posts from this date (ISO 8601)
endDate	string	No	Filter posts until this date (ISO 8601)
latestCount	number	No	Return only the latest X posts (max 100)

list_feedback_widgets

List all your feedback widgets (id, name, active status, creation date).

get_feedback_responses

Get responses for a specific feedback widget.

Parameter	Type	Required	Description
feedbackWidgetId	number	Yes	The ID of the feedback widget
startDate	string	No	Filter responses from this date (ISO 8601)
endDate	string	No	Filter responses until this date (ISO 8601)
latestCount	number	No	Return only the latest X responses (max 100)

list_roadmaps

List all your roadmaps with their sections (id, name, public name, active status, sections).

get_roadmap

Get a full roadmap with all sections and items. Supports search and tag filtering within the roadmap.

Parameter	Type	Required	Description
roadmapId	number	Yes	The ID of the roadmap
startDate	string	No	Filter items updated from this date (ISO 8601)
endDate	string	No	Filter items updated until this date (ISO 8601)
latestCount	number	No	Return latest X items per section (max 100)
query	string	No	Case-insensitive substring match on item name or description
tagNames	string[]	No	Filter items whose tag name matches any of these (case-insensitive exact match)

list_nps_widgets

List all your NPS widgets (id, name, active status, creation date).

get_nps_score

Get the NPS score and response breakdown for an NPS widget. NPS = % promoters (rating 9-10) − % detractors (rating 0-6); passives are 7-8.

Parameter	Type	Required	Description
npsWidgetId	number	Yes	The ID of the NPS widget
startDate	string	No	Filter responses from this date (ISO 8601)
endDate	string	No	Filter responses until this date (ISO 8601)

Returns npsScore, totalResponses, and counts/percentages for promoters, passives, and detractors.

get_nps_responses

Get individual responses for an NPS widget.

Parameter	Type	Required	Description
npsWidgetId	number	Yes	The ID of the NPS widget
startDate	string	No	Filter responses from this date (ISO 8601)
endDate	string	No	Filter responses until this date (ISO 8601)
latestCount	number	No	Return only the latest X responses (max 100)

get_widget_stats

Get event counts and unique users per widget for a given widget type. Returns one entry per widget with a breakdown by event type. Useful for comparing widgets of the same type (e.g. "which tours have the best completion rate?") or inspecting a single widget when entityId is provided.

Parameter	Type	Required	Description
entityType	enum	Yes	One of tour, checklist, smartTip, announcement, changelog, npsWidget, roadmap
entityId	number	No	Filter to a single widget
startDate	string	No	Start date (ISO 8601). Defaults to 30 days ago
endDate	string	No	End date (ISO 8601). Defaults to today

Max range 90 days. uniqueUsers is a per-day sum approximation — it may double-count users active across multiple days.

get_stats_summary

Get a company-wide stats summary across all widget types. Returns primary and secondary metrics (completion rate, open rate, action rate), event counts, period-over-period comparison, and a trend series for each widget type.

Parameter	Type	Required	Description
startDate	string	No	Start date (ISO 8601). Defaults to 7 days ago
endDate	string	No	End date (ISO 8601). Defaults to today

The trend series is daily when the range is 21 days or less, weekly otherwise. Max range 90 days.

list_tags

List all tags in the company (id, name, colors). Useful for looking up tag names to pass to create_changelog_post and update_changelog_post.

create_changelog_post

Create a new changelog post. Posts default to draft (active: false) unless active: true is explicitly passed.

Parameter	Type	Required	Description
changelogId	number	Yes	The ID of the changelog to add the post to
title	string	Yes	Post title
description	string	No	Post description as HTML (e.g. <p>, <ul>, <strong>)
date	string	No	Post date (ISO 8601). Defaults to now
active	boolean	No	Whether the post is live. Defaults to false
tagNames	string[]	No	Tag names to attach. Must match existing tags (case-insensitive)

Unknown tagNames return an error listing the unresolved names — use list_tags first.

update_changelog_post

Update an existing changelog post. Only the fields you provide are changed.

Parameter	Type	Required	Description
changelogItemId	number	Yes	The ID of the post to update
title	string	No	New title
description	string	No	New HTML description
date	string	No	New post date (ISO 8601)
active	boolean	No	Whether the post is live
tagNames	string[]	No	Replacement tag names. Omit to leave tags unchanged; pass [] to clear

Example usage

Once connected, you can ask your AI assistant things like:

• "What are our latest changelog posts?"
• "Show me feedback from the last week"
• "What items are on our product roadmap?"
• "Search the roadmap for items about 'integration' tagged as 'New Feature'"
• "What's our NPS score this quarter, and what are the detractors saying?"
• "Which product tour has the best completion rate over the last month?"
• "How did our overall engagement change this week vs last week?"
• "Summarize the feedback we got this month"
• "Draft a changelog post about the webhook improvements we shipped this week, tag it as Improvement, leave it as a draft"