JetBrains IDEs
Connect Runstack MCP in IntelliJ, PyCharm, WebStorm and other JetBrains IDEs via the MCP Server plugin.
Source: https://runstack.engineer/docs/mcp-jetbrains
jetbrains
JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, GoLand, Rider, etc.) can call Runstack's MCP tools through the MCP Server integration. The same https://app.runstack.engineer/mcp endpoint is used (Streamable HTTP, Authorization: Bearer rnsk_…).
Prerequisites
- JetBrains IDE 2025.1+ (2025.2+ has integrated MCP Server; earlier versions use the plugin)
- Runstack API key — API Keys
Install plugin (if needed)
- Settings → Plugins → Marketplace → search MCP Server → Install → Restart IDE.
- On 2025.2+, the server is built-in under Settings → Tools → MCP Server.
Configuration
Option A — IDE settings table
- Go to Settings → Tools → AI Assistant → Model Context Protocol (MCP) or Settings → Tools → MCP Server.
- Click Add → paste a JSON entry for Runstack.
HTTP (hosted /mcp):
{
"mcpServers": {
"runstack": {
"url": "https://app.runstack.engineer/mcp",
"headers": {
"Authorization": "Bearer rnsk_YOUR_API_KEY"
}
}
}
}
If your build shows serverUrl instead of url, use serverUrl.
stdio (@rnsk/mcp) — alternative:
{
"mcpServers": {
"runstack": {
"command": "npx",
"args": ["-y", "@rnsk/mcp"],
"env": {
"RUNSTACK_API_KEY": "rnsk_YOUR_API_KEY"
}
}
}
}
- Apply → restart the IDE (or the AI Assistant) if tools do not appear.
Option B — file config (some JetBrains MCP builds)
Some installations read ~/.config/jetbrains/mcp.json or import from Claude Desktop (claude_desktop_config.json). When available, you can paste the same mcpServers.runstack block there and use Import from Claude in the settings UI.
Verify
Open the JetBrains AI Assistant chat and prompt: “Use Runstack to search for Slack tools”. Confirm search_tools, check_authentication, initiate_connection, execute_tool are listed as available tools. OAuth for Gmail/GitHub still opens via Runstack when check_authentication returns token_missing.