AI Agents Built My Product Plan! (Google ADK - PRD and product lifecycle Automation)

Welcome to the Idea-to-Blueprint-Pipeline (ITBP) Agent project! This repository showcases a powerful AI agent system built using Google's Agent Development Kit (ADK) that transforms product ideas into comprehensive project blueprints through a structured, phase-based workflow.

This project serves as a practical example and is featured in a YouTube video titled "AI Agents Built My Product Plan! (Google ADK - PRD and product lifecycle Automation)" on the @TonyAlfredsson YouTube channel. Follow along to see how it was built and how it can be extended!

✨ Introduction

This agent system demonstrates how to create a comprehensive product development pipeline using ADK-based agents. It provides a structured workflow that guides users through the entire process of transforming an idea into a detailed implementation plan:

• Brainstorm and refine initial product concepts
• Conduct market and technical research
• Create structured project briefs
• Define detailed product requirements and epics
• Design technical architecture
• Validate the plan from a product perspective
• Generate detailed user stories for implementation

Built entirely with the Google ADK framework, it uses a team of specialist agents (Analyst, PM, Architect, PO/SM) coordinated by a central orchestrator to guide users through each phase of the product development lifecycle.

🚀 Features

• Idea Exploration: Brainstorm and refine product concepts with AI assistance
• Market Research: Conduct research on concepts and market opportunities
• Project Brief Creation: Transform ideas into structured project briefs
• PRD Generation: Create comprehensive Product Requirements Documents
• Epic Definition: Structure work into logical epics with clear requirements
• Architecture Design: Design technical architecture based on requirements
• Plan Validation: Validate the plan from a product perspective
• User Story Generation: Create detailed user stories for implementation
• ADK Integration: Built entirely within the Google Agent Development Kit framework
• Phase-Based Workflow: Structured process with clear transitions between phases
• Specialist Agents: Team of expert agents with specific roles and responsibilities

🛠️ How It Works

The itbp_agent acts as an orchestrator, coordinating a team of specialist agents through a phase-based workflow:

1. Analyst Agent: Facilitates brainstorming, conducts research, and creates project briefs
2. PM Agent: Defines product requirements and structures work into epics
3. Architect Agent: Designs technical architecture based on requirements
4. PO/SM Agent: Validates the plan and generates user stories

The workflow follows these phases:

1. GET_IDEA - Collect the initial product/project idea from the user
2. ANALYST_BRAINSTORM - Explore and refine the idea
3. ANALYST_RESEARCH - Conduct market and technical research
4. ANALYST_BRIEF - Create a structured Project Brief
5. PM_DEFINE - Define product requirements and structure work into Epics
6. ARCHITECT_DESIGN - Design technical architecture
7. POSM_VALIDATE - Validate the plan from a product perspective
8. POSM_STORIES - Generate detailed user stories
9. FINISHED - Complete the blueprint process

💻 Technology Stack

• Framework: Google Agent Development Kit (ADK)
• Language: Python 3.x
• LLMs: Google Gemini 2.5 Pro (via google-genai library)
• Core Libraries: google-adk, google-genai
• Package Management: uv (recommended) or pip

⚙️ Setup & Installation

Follow these steps to get the agent running on your local machine.

1. Clone the Repository

git clone https://github.com/LiquidError/adk-from-idea-to-blueprint-agents
cd adk-from-idea-to-blueprint-agents

2. Configure API Keys

This project requires API keys for Google (for agent control and LLM access) and Tavily (for the deep research tool).

• Copy the example environment file:

  cp .env.example .env

• Edit the .env file and add your actual API keys:

  GOOGLE_API_KEY="YOUR_GOOGLE_API_KEY_HERE"
  TAVILY_API_KEY="YOUR_TAVILY_API_KEY_HERE"

• You can get a Google API key from Google Cloud Console or Goggle AI Studio

• You can get a Tavily API key from Tavily's documentation

3. Set Up a Virtual Environment

It's highly recommended to use a virtual environment.

• Using uv (Recommended)
  • Install uv if needed (see official uv documentation).
  • Create and activate the virtual environment:

    uv venv
    source .venv/bin/activate # (or .venv\Scripts\activate on Windows)

• Using Python's built-in venv
  • Create and activate the virtual environment:

    # On macOS/Linux:
    python3 -m venv .venv && source .venv/bin/activate
    # On Windows:
    python -m venv .venv && .venv\Scripts\activate

4. Install Dependencies

• Using uv (Recommended):

  uv sync

• Using pip:

  pip install -r requirements.txt

▶️ Running the Agent

You can interact with the agent using the ADK CLI:

1. Interactive Web UI (adk web)

This is the easiest way to test.

adk web --port=8080 ./v1/agents/

Navigate to http://localhost:8080 in your browser. Select the itbp_agent. You can then:

• Start by providing your product idea
• Follow the guided workflow through each phase
• Review and provide feedback on outputs at each stage
• Receive a comprehensive project blueprint at the end

⚠️ Known Limitations

• No Artifact Storage: This version does not produce persistent artifacts. All output text needs to be copied and pasted into another document for preservation.
• API Timeouts: Some timeouts in long and complex queries might occur due to rate limits or an overloaded API, especially when using the deep research tool.
• Research Tool Dependencies: The deep research functionality requires a valid Tavily API key to work properly.

📺 YouTube Video & Inspiration

This project is featured in a YouTube video titled "AI Agents Built My Product Plan! (Google ADK - PRD and product lifecycle Automation)".

• Watch: youtube.com/@TonyAlfredsson
• Inspired by:
  • BMadCode - Innovative AI methodologies
  • BMAD METHOD - Structured approach to AI development
  • AI with Brandon - Practical AI applications

🔗 References

• Google Agent Development Kit (ADK) Documentation: google.github.io/adk-docs
• UV Documentation: astral.sh/uv

📜 License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.

🤝 Contributing & Feedback

Contributions are welcome! Please feel free to open an issue or submit a pull request.

Feedback and suggestions are also encouraged! Please leave comments on the YouTube videos or open an issue in this repository.