What Is the Model Context Protocol?
The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external tools and data sources. With the FeaturePulse MCP server, your AI assistant gets direct access to your feature request data — no copy-pasting, no context switching.
Available Tools
- list_feature_requests — Browse and filter all feature requests with full MRR data
- get_project_stats — High-level overview: total requests, votes, and MRR by status/priority
- search_feedback — Full-text search across feature requests
- analyze_feedback_by_group — Group requests by status or priority with aggregated MRR
- update_feature_status — Change status or priority of a feature request
Setup with Codex and Claude
For Claude Code, run this single command:
claude mcp add --transport http featurepulse https://featurepul.se/api/mcpIn Claude Code, open /mcp and authenticate FeaturePulse. For Codex, run:
codex mcp add featurepulse --url https://featurepul.se/api/mcp
codex mcp login featurepulseFor Codex roadmap updates, use codex mcp login featurepulse --scopes feedback:read,feedback:write and approve updates in the browser. In the Codex app or Claude connector settings, add the same remote MCP URL. Manage and revoke access in Connected apps.
Setup with Cursor / Windsurf
Add https://featurepul.se/api/mcp as a remote MCP server in an editor that supports Streamable HTTP and OAuth, then follow browser login to approve access.
Example Prompts
- "What are the top 5 feature requests by MRR impact?"
- "Show me all approved requests that aren't started yet"
- "How much MRR is behind high-priority pending requests?"
- "Search for feedback about dark mode"
- "Mark the dark mode request as in_progress"
Pair the MCP integration with GitHub integration for a powerful AI-driven workflow: analyze feature requests with Claude, then automatically create GitHub Issues for the ones you decide to build.