CogniReq: The First FOSS AI-Native Requirements Management Platform

Empowering Requirements Engineers with Intelligent Interview Guidance and Structured Artifact Generation.

Introduction

CogniReq is a Free and Open Source Software (FOSS) platform dedicated to revolutionizing Requirements Engineering through AI-native solutions. The mission is to provide innovative tools that enhance clarity, precision, and efficiency in the requirements gathering process.

Currently, CogniReq's foundational feature is the Interview Guide Assist (IGA), a specialized AI tool designed to improve the preparation for critical stakeholder interviews. The IGA addresses the common challenge of inconsistent or incomplete requirements gathering by offering a highly structured and systematic approach, moving beyond traditional, single-prompt LLM interactions.

Key Feature: Interview Guide Assist (IGA)

The Problem

The Interview Guide Assist (IGA) tackles the critical issue of inconsistent and incomplete requirements gathering, often stemming from a lack of standardized, rigorous guidelines for interview preparation in Requirements Engineering literature. This gap may lead to crucial information being missed during stakeholder interviews.

The Solution: IGA's Approach

The IGA provides a highly structured and systematic approach to interview preparation. It moves beyond simple, single-prompt LLM interactions by requiring the Requirements Engineer (RE) to input initial context (project name, summaries of the project and stakeholder). This is followed by an iterative, conversational refinement process. The ultimate goal is to produce a well-structured Interview Guide Artifact that minimizes the risk of missing crucial information during the actual interview.

The AI assistant functions as a structured co-pilot and expert interpreter. It leverages modern LLM capabilities, specifically JSON Schema Enforcement (Structured Output), to guarantee the integrity and reliability of the Interview Guide Artifact.

User Experience

See for yourself with a small demo GIF:

---

---

In the main user interface, the RE interacts with the AI via a chat pane (1/2 width) to issue refinement commands, such as "Make Question 5 more open-ended" or "Refine the Project Summary." The AI receives the entire current state of the structured artifact along with the command and returns a complete, updated version of the guide. This updated content instantly replaces the previous version in the main display pane (1/2 width). This mechanism ensures seamless synchronization and allows the AI to provide expert-level guidance, not just text generation, while the RE retains final editorial control.

A unique Cite Feature allows the user to click any part of the structured guide, which automatically initiates a refinement command in the chat. This tightly coupled, dual-pane system enables the RE to efficiently transform raw project knowledge into a highly systematic and professional interview script, significantly enhancing the rigor of the requirements elicitation process.

Key Artifact Components

The AI must generate and maintain the following key features within the Interview Guide Artifact:

• Interview Goal: A clear objective for the interview.
• General Interview Advice: Best practices and tips for the RE.
• Questions: A list of specific, editable questions.
• Note: Each question is paired with a corresponding note, providing rationale or follow-up advice for the RE.

Technology Stack

CogniReq is built with a modern, full-stack approach, leveraging robust and efficient technologies:

• Backend: Go
• Frontend: Vue.js 3 (Composition API), TypeScript, TailwindCSS v4
• LLM Integration: Google Gemini
• Quality Assurance:
  • ✅ >70% Test Coverage, including mocked AI behavior for various interactions (create, update, delete questions, refine context data).
  • ✅ Automatic Testing using GitHub Actions ensures continuous integration and code quality.

Architecture Overview

The application employs a strong separation of concerns, following a layered architecture to ensure maintainability, scalability, and clarity:

• Web Handler: Manages request/response DTOs.
• Service Layer: Contains the core business logic and domain models.
• Database Layer: Handles data persistence, utilizing SQLC for efficient and type-safe database interactions.

Getting Started (How to Run)

To get CogniReq up and running quickly, it's recommended to use Docker and Docker Compose.

Prerequisites

• Docker: Ensure Docker is installed and running on your system. You can download it from Docker's official website.
• Docker Compose: This is usually included with Docker Desktop installations, see Docker's official website for more info.

Steps

1. Clone the Repository:

   git clone git@github.com:jensilo/cognireq.git
   cd cognireq

2. Configure Environment Variables: Navigate to the docker/dev directory and copy the example environment file:

   cd docker/dev
   cp .env.dist .env

   Edit the .env file and fill in your Google Gemini API key:

   # .env
   GEMINI_API_KEY = YOUR_GEMINI_API_KEY_HERE
   

   (You can obtain a Google Gemini API key from Google AI Studio.)

3. Build and Run with Docker Compose: From the docker/dev directory, run:

   docker compose up

   This command will:

   • Build the Docker images for the backend and database.
   • Start the database, backend, and frontend services.
   • The application will be accessible at http://localhost:5173 (frontend) and the API at http://localhost:8080 ( backend).

4. Access the Application: Open your web browser and navigate to http://localhost:5173.

5. Stop the Application: To stop the running services, press Ctrl+C in the terminal where docker compose is running, or execute:

   docker compose down

Roadmap and Future Vision

CogniReq is envisioned as a comprehensive AI-native Requirements Management platform. While the Interview Guide Assist (IGA) is its initial core offering, our roadmap includes expanding its capabilities to cover the full requirements lifecycle. Future features may include:

• Convenient Requirements Generation: Conveniently generate high-quality requirements directly and seamlessly from requirements elicitation interview results.
• Advanced Requirements Analysis: AI-driven insights into requirements quality, consistency, and completeness.
• Automated Traceability: Tools for establishing and maintaining traceability between requirements, design, and tests.
• Integration with RE Tools: Seamless integration with popular requirements management and project planning tools.
• Collaborative Features: Enhancing teamwork in requirements elicitation and management.

The goal is to continuously evolve CogniReq, making it the go-to FOSS solution for intelligent requirements engineering. (maybe one day... ;))

Contributing

Contributions are welcome from anybody! Just submit a PR, or reach out to me.

Disclaimer

CogniReq is currently under active development and is considered a prototype. Production-grade stability or security is not guranteed. Users should exercise caution and not rely on it for mission-critical applications without thorough independent validation.

License

This project is licensed under the MIT License.