中文 | English
AI-Powered Requirements Analysis & Management System
In enterprise digital transformation and agile delivery projects, requirements management is often the biggest bottleneck:
- 📧 Requirements are scattered across emails, meeting notes, chat messages, and documents
- ⏰ Writing structured requirements takes hours and quality varies by person
- 🔄 Manual synchronization between requirements documents and project management tools (Jira, Azure DevOps) is error-prone
- 🌐 Language barriers slow down global team collaboration
- 🔒 Compliance concerns prevent many enterprises from adopting public cloud AI tools
RequirementTracker is an AI-powered multi-agent system that automatically transforms informal requirement descriptions into professional, structured documents and seamlessly integrates with enterprise collaboration platforms.
- ⚡ 8x faster - Reduce requirements documentation time from hours to minutes
- 📉 60% less manual work - Automate requirement extraction, formatting, and tool integration
- ✅ 40% fewer missed requirements - AI-powered analysis ensures completeness
- 🔐 Enterprise-ready - Support for on-premises LLMs (Qwen), Azure OpenAI, and custom models
Powered by crewAI, two specialized agents work together:
- Requirements Analyst Agent: Extracts and structures key information from unstructured text
- Document Formatter Agent: Generates professional Markdown documentation with enterprise-standard formats
- ✅ Qwen (Alibaba Cloud) - Best for Chinese requirements, China compliance-ready
- ✅ Azure OpenAI - Enterprise-grade, Microsoft 365 integrated
- ✅ Grok (xAI) - Cost-effective, innovative alternative
- ✅ Custom Models - Bring your own LLM (on-premises or cloud)
- Azure DevOps API: Automatically create Work Items with structured requirements
- Confluence API: Generate and publish documentation pages
- Extensible Architecture: Easy to add Jira, Notion, Monday.com, etc.
- CLI: Perfect for automation, CI/CD pipelines, and power users
- Streamlit Web UI: User-friendly interface for non-technical stakeholders
- 87% unit test coverage
- Comprehensive CI/CD pipeline
- Modular, extensible architecture
- Environment-based configuration management
┌─────────────────────────────────────────────────────────┐
│ User Interface Layer │
│ CLI (Command Line) │ Web UI (Streamlit) │
└────────────────────┬────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ Multi-Agent Processing Layer │
│ ┌──────────────────────────────────────────────────┐ │
│ │ Requirements Analyst → Document Formatter │ │
│ │ Agent Agent │ │
│ └──────────────────────────────────────────────────┘ │
└────────────────────┬────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ LLM Integration Layer │
│ Qwen │ Azure OpenAI │ Grok │ Custom Models │
└────────────────────┬────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ External Integrations Layer │
│ Azure DevOps │ Confluence │ Jira (Coming Soon) │
└─────────────────────────────────────────────────────────┘
- Python 3.11 or higher
- API keys for your chosen LLM provider (Qwen, Azure OpenAI, or Grok)
- (Optional) Azure DevOps / Confluence credentials for integration
# Clone the repository
git clone https://github.com/loootte/req_agent.git
cd req_agent
# Install dependencies
pip install -r requirements.txtCreate a .env file in the project root:
# LLM Configuration (choose one)
QWEN_API_KEY=your_qwen_api_key
QWEN_MODEL_NAME=qwen-max
# Or use Azure OpenAI
# AZURE_OPENAI_API_KEY=your_key
# AZURE_OPENAI_ENDPOINT=your_endpoint
# AZURE_OPENAI_DEPLOYMENT_NAME=your_deployment
# Or use Grok
# GROK_API_KEY=your_grok_key
# Optional: Tool Integration
# AZURE_DEVOPS_PAT=your_pat
# AZURE_DEVOPS_ORG=your_organization
# CONFLUENCE_API_TOKEN=your_token# Analyze requirements from a text file
python main.py --input requirements.txt --model qwen
# With Azure DevOps integration
python main.py --input requirements.txt --model azure --create-workitem
# With Confluence integration
python main.py --input requirements.txt --model qwen --create-confluence# Launch Streamlit interface
python -m streamlit run src/requirement_tracker/webapp.py
# Open browser at http://localhost:8501我们需要一个动态定价系统，能够根据市场行情自动调整铁矿石的报价。
系统应该每天早上8点自动运行，从外部数据源获取最新价格，
然后根据我们的定价策略计算出新的报价，并发送给销售团队审核。
## Requirements Analysis: Dynamic Pricing System
### Problem Statement
Need an automated dynamic pricing system for iron ore that adjusts
quotations based on real-time market conditions.
### Business Goals
- Improve pricing responsiveness to market changes
- Reduce manual pricing work for sales team
- Maintain competitive pricing strategy
### Key Deliverables
1. Automated daily pricing calculation engine
2. External market data integration
3. Pricing strategy configuration module
4. Sales team review and approval workflow
### Acceptance Criteria
- System runs daily at 8:00 AM automatically
- Successfully retrieves latest market prices from external sources
- Calculates new quotations based on configurable pricing rules
- Sends results to sales team for review via email/dashboard
### Technical Considerations
- Data source reliability and latency
- Pricing algorithm transparency and auditability
- Security for pricing strategy parameters
### Risk Assessment
- External data source availability (Medium risk)
- Pricing algorithm accuracy validation (High priority)
- Change management for sales team adoption (Medium risk)- ✅ Azure DevOps Work Item created with structured requirements
- ✅ Confluence Page published for team collaboration
- ✅ Email notification sent to stakeholders (coming soon)
- Accelerate client project delivery
- Standardize requirements documentation across teams
- Reduce billable hours spent on documentation
- Improve requirements quality and consistency
- Bridge communication gaps between business and IT
- Enable faster digital transformation initiatives
- Automate backlog grooming and user story creation
- Reduce time spent in requirements clarification meetings
- Improve Jira/ADO hygiene automatically
- Break language barriers with multi-language support
- Ensure consistent requirements format across regions
- Reduce asynchronous communication overhead
- Multi-agent requirements analysis system
- Support for Qwen, Azure OpenAI, Grok
- CLI and Web UI interfaces
- Azure DevOps and Confluence integration
- Jira integration
- Notion integration
- Monday.com integration
- Email notifications
- Slack/Teams integration
- Batch processing for multiple requirements
- Requirements template library (by industry/scenario)
- Version control and change tracking
- Conflict and dependency detection
- Requirements prioritization scoring
- Multi-language UI (English, Chinese, Japanese)
- Multi-tenant SaaS platform
- Role-based access control (RBAC)
- Analytics and reporting dashboard
- Custom workflow automation
- Enterprise SSO integration
- On-premises deployment support
# Run all tests
pytest
# Run with coverage report
pytest --cov=src --cov-report=html
# View coverage report
open htmlcov/index.htmlCurrent test coverage: 87%
We welcome contributions! Here's how you can help:
- 🐛 Report bugs via GitHub Issues
- 💡 Suggest features or improvements
- 📝 Improve documentation (especially use cases and examples)
- 🔧 Submit pull requests for bug fixes or new features
- 🌐 Add LLM integrations (Claude, Gemini, etc.)
- 🔗 Add tool integrations (Jira, Notion, Linear, etc.)
# Fork and clone your fork
git clone https://github.com/loootte/req_agent
# Create a feature branch
git checkout -b feature/your-feature-name
# Make changes and test
pytest
# Submit pull requestInterested in enterprise deployment, custom features, or consulting services?
- 📧 Email: [email protected]
- 📞 Phone: +86 180 5718 8110
We offer:
- Custom LLM integration (private/on-premises models)
- Enterprise tool integration (SAP, Oracle, Salesforce, etc.)
- Training and workshops for your teams
- Dedicated support and SLA
- White-label deployment options
This project is licensed under the MIT License - see the LICENSE file for details.
- crewAI - Excellent multi-agent framework
- Streamlit - Beautiful UI framework
- LangChain - Inspiration for agent design
- All contributors and pilot project teams
If you find this project useful, please consider giving it a star! ⭐
- GitHub Issues: Report bugs or request features
Made with ❤️ by Zhong Le
Transforming requirements management with AI, one document at a time.