18 AI product-management robots for market research, financials, and execution specs. No API keys
AutoPM is an agentic MCP server that transforms raw business ideas into complete product strategies. It orchestrates a specialized AI team to handle real-time market research, competitive deep-dives, persona building, and dynamic financial modeling, culminating in a fully-styled, pitch-ready presentation.
🟢 New to this? Check out the Quick Start Guide for Beginners for step-by-step setup instructions!
Real-world Product Management requires distinguishing between finding the right thing to build and defining how to build it right. AutoPM mirrors this exact discipline through a strictly gated, two-phased workflow:
By enforcing Phase 1 as a prerequisite for Phase 2, AutoPM prevents the most common PM trap: rushing into solution design and feature-building without a validated market strategy.
claude_desktop_config.json).command: "node" and args: ["/absolute/path/to/AutoPM/src/mcp-server.js"].interview tool, iterate via run-robot, and export using generate-presentation!Because AutoPM is built on the standard Model Context Protocol (MCP), you can use it with any compatible client!
~/Library/Application Support/OpenAI/ChatGPT/mcp.json and add the productflow server configuration similar to Claude Desktop (command: node, args: ["/absolute/path/to/AutoPM/src/mcp-server.js"]). Restart the app.command, Command: node, Args: /absolute/path/to/AutoPM/src/mcp-server.js."mcpServers": {
"productflow": {
"command": "node",
"args": ["/absolute/path/to/AutoPM/src/mcp-server.js"]
}
}
npm installrun.js to change your target product ideanode run.jssrc/ - Core MCP Server implementation (mcp-server.js)robots/ - The specialized autonomous agent logicbrain/ - Persistent memory, agent instructions, and user feedback engineleader/ - TeamLeader orchestrator managing the flowutils/ - Output engines and file toolingplans/ - Auto-generated HTML presentation deliverablesThe ProductFlow Cockpit UI supports full URL routing, browser history (Back/Forward navigation), page-refresh persistence, and deep-linking via a hash-based routing system.
autopm.ai)Instead of accessing the Cockpit UI via http://localhost:4321, you can configure your machine to access it via a custom local domain: http://autopm.ai.
sudo nano /etc/hosts
127.0.0.1 autopm.ai
Ctrl + O, Enter, then Ctrl + X).You have two options for starting the HTTP Cockpit server:
Option A: Run on Port 4321 (Default) Start the server normally:
npm run http
Access it in the browser at: http://autopm.ai:4321
Option B: Run on Port 80 (No port number in URL)
Port 80 is a privileged port, so you must start Node with sudo permissions:
sudo PRODUCTFLOW_HTTP_PORT=80 node src/http-server.js
Access it in the browser directly at: http://autopm.ai
Source-derived launch command. Check the maintainer’s required arguments and credentials before running:
npx -y @productflow/autopmMerge this template into ~/Library/Application Support/Claude/claude_desktop_config.json. Keep existing servers. Add any arguments, credentials, and permissions required by the maintainer; this template has not been install-tested.
{
"mcpServers": {
"io-github-fluidumber-autopm": {
"command": "npx",
"args": [
"-y",
"@productflow/autopm"
]
}
}
}Restart Claude Desktop completely for changes to take effect. Confirm the server appears connected in the client’s tool list, then try a read-only example from its documentation.
Claude Desktop setup reference@productflow/autopmnpmio.github.fluidumber/autopm works with any MCP-compatible client. Copy the config snippet from the Configuration section above and add it to the file shown for your client, then restart the application.
~/Library/Application Support/Claude/claude_desktop_config.jsonRestart Claude Desktop completely for changes to take effect.~/.cursor/mcp.jsonRestart Cursor for changes to take effect..vscode/mcp.jsonReload VS Code window for changes to take effect.~/.codeium/windsurf/mcp_config.jsonRestart Windsurf for changes to take effect..mcp.jsonSave at the project root, then start Claude Code in that project and review the MCP server approval prompt. Keep real credentials out of shared files.