StrataBRD Pro - AI Business Requirements Architect

StrataBRD Pro is an AI agent designed to autonomously generate comprehensive, industry-standard Business Requirements Documents (BRDs) from high-level concepts or concise prompts. This initial version focuses on core functionalities using the LangGraph framework.

Overview

The agent embodies the persona of "StrataBRD Pro," an elite AI Business Requirements Architect. It leverages a Large Language Model (LLM) to understand user input and generate structured BRD sections. The underlying architecture is built with LangGraph, allowing for a stateful, multi-step process.

Current Capabilities (Initial Version):

• Accepts a high-level concept from the user via a command-line interface (CLI).
• Generates an initial BRD draft in Markdown format, including:
  • Executive Summary
  • Vision & Scope
  • Basic Functional Requirements (as user stories with acceptance criteria)
• Follows the core principles and output standards defined in its persona.
• Interactive Clarification Loop: If initial input is insufficient, the agent can ask clarifying questions to refine its understanding before generating the BRD. This loop is configurable (e.g., max rounds).
• Project Persistence: Automatically saves your work (including conversation history and current understanding) and allows you to load and continue existing projects. Project data is stored locally in the brd/saved_projects/ directory.

Project Structure

• brd/: Main package for the agent.
  • prompts.py: Contains the detailed persona and system prompts for the LLM.
  • agent.py: Core logic for BRD generation and clarification, including LLM interaction.
  • graph.py: Defines the agent's stateful workflow using LangGraph (state, nodes, edges), managing the clarification and generation process.
  • persistence.py: Handles saving and loading project states to/from JSON files.
• main.py: CLI entry point to interact with the agent, managing project creation, loading, and user interaction with the graph.
• tests/: Unit tests for various components (agent logic, graph flow, persistence).
• requirements.txt: Python dependencies.
• run_tests.sh: Script to execute unit tests.

Setup

1. Clone the repository (if applicable).

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Linux/macOS
   # venv\Scripts\activate    # On Windows

3. Install dependencies:

   pip install -r requirements.txt

4. Set up OpenAI API Key: The agent uses OpenAI's GPT models. You need to provide your API key.

   • Create a file named .env in the project root directory (next to main.py).
   • Add your API key to the .env file like this:

     OPENAI_API_KEY='your_actual_openai_api_key_here'
     

   Important: Keep your API key confidential. Do not commit the .env file to version control if this were a shared repository.

Configuration

Beyond the API key, you can configure the LLM behavior using environment variables (also in the .env file or set globally):

• OPENAI_API_KEY (Required): Your OpenAI API key.
• OPENAI_MODEL_NAME (Optional): Specifies the OpenAI model to use.
  • Default: "gpt-3.5-turbo"
  • Example: OPENAI_MODEL_NAME="gpt-4-turbo-preview"
• OPENAI_TEMPERATURE (Optional): Controls the creativity of the LLM's responses. It's a float value.
  • Default: 0.7
  • Example: OPENAI_TEMPERATURE="0.5"

The agent's clarification question mechanism is designed to receive questions from the LLM in a JSON-formatted list of strings. This is primarily an internal detail relevant for prompt engineering or contributions.

How to Run

1. Ensure you have completed the Setup and Configuration steps above (especially setting OPENAI_API_KEY).
2. Run the CLI from the project root directory:

   python main.py

3. CLI Interaction:
   • The agent will initialize and display a welcome message.
   • You will be presented with a Main Menu:
     • (N) New Project: Start a new BRD. You'll be asked for a project name and then your initial concept.
     • (L) Load Project: Load a previously saved project. A list of available projects will be shown for you to choose from.
     • (E) Exit: Quit the application.
   • If you load a project that was previously waiting for your answers to clarification questions, these questions will be re-displayed.
   • After the agent processes your input (either generating a BRD or asking clarification questions), your project state is automatically saved.
   • If a BRD or set of questions is generated, you'll be asked if you want to continue with the current project (e.g., provide answers, refine further), start a new one, load another, or exit.

Project Persistence

StrataBRD Pro supports project persistence. When you start a new BRD, you'll be asked to name your project (or one will be auto-generated if you leave it blank). All progress, including your inputs, the agent's generated content, conversation history, and current understanding, is automatically saved to a JSON file in the brd/saved_projects/ directory. When you run main.py, the main menu allows you to load an existing project or start a new one.

Running Tests

1. Ensure dependencies are installed (see Setup).
2. Make the test script executable (if not already):

   chmod +x run_tests.sh

3. Run the tests from the project root directory:

   ./run_tests.sh

   The test suite covers agent logic, graph workflow, and persistence mechanisms. Mocks are used to prevent actual LLM API calls during tests.

Future Enhancements (Planned)

• Implement all 14 sections of the BRD structure (currently focuses on initial key sections).
• Integrate a proper Retrieval Augmented Generation (RAG) system for domain-specific knowledge.
• Implement multi-stage validation.
• And much more as per the original issue specification...