An AI-powered multi-agent system that generates technology trend reports. Enter any technology (Artificial Intelligence, Blockchain, Quantum Computing) and get a professional Word document with market analysis, key players, use cases, and strategic outlook.
Charts are generated from the analysis output and embedded directly into the DOCX report. Examples from the bundled Artificial Intelligence mock data:
Key Player Market Relevance: horizontal bar chart ranking leading companies
Use Cases by Impact Level: distribution of identified use cases across impact tiers
CLI Input ("Quantum Computing")
│
▼
┌─────────────────────┐
│ Orchestrator │ Coordinates the 3-phase pipeline
└────────┬────────────┘
│
┌────┴─────┐
▼ ▼
┌────────┐ ┌──────────┐
│Research│ │ Analysis │
│ Agent │ │ Agent │
└────┬───┘ └────┬─────┘
│ │
▼ ▼
┌─────────────────────┐
│ Report Generator │ → DOCX Output
└─────────────────────┘
| Agent | Role | Technology |
|---|---|---|
| ResearchAgent | Gathers live web intelligence on the technology | Server-side web search, adaptive thinking |
| AnalysisAgent | Extracts structured insights from raw research | Structured outputs (Pydantic), adaptive thinking |
| ReportGenerator | Produces a professional multi-page DOCX report | python-docx with color-coded tables |
# 1. Clone the repo
git clone https://github.com/eugen-goebel/tech-trend-agent.git
cd tech-trend-agent
# 2. Install dependencies
pip install -r requirements.txt
# 3a. Test without an API key (uses sample AI data)
python main.py --dry-run
# 3b. Full run with Anthropic API key
echo "ANTHROPIC_API_KEY=sk-ant-..." > .env
python main.py "Quantum Computing"
python main.py "Blockchain"
python main.py "Edge Computing"The report is saved to ./output/tech_trend_<technology>_<date>.docx.
# Run the full test suite (52 tests, no API key needed)
python -m pytest tests/ -vThe test suite covers:
- Model validation: Pydantic schemas, Literal constraints, serialization
- Mock data integrity: ensures dry-run data is complete and valid
- Report generation: DOCX output, section presence, table structure
- Agent logic: web search tool usage, pause_turn handling, structured outputs
- CLI integration: argument parsing, dry-run mode, error handling
Running python main.py "Artificial Intelligence" produces a ~10-page Word document:
tech-trend-agent/
├── main.py # CLI entry point (supports --dry-run)
├── agents/
│ ├── researcher.py # Web search intelligence gathering
│ ├── analyst.py # Structured analysis (Pydantic models)
│ ├── orchestrator.py # Pipeline coordinator
│ └── mock_data.py # AI sample data for --dry-run mode
├── utils/
│ └── report_generator.py # Professional DOCX generation
├── tests/
│ ├── test_models.py # Pydantic model validation tests
│ ├── test_mock_data.py # Mock data integrity tests
│ ├── test_report_generator.py # DOCX generation tests
│ ├── test_agents.py # Agent logic tests (mocked API)
│ └── test_cli.py # CLI integration tests
├── output/ # Generated reports (git-ignored)
├── requirements.txt
└── .env.example
| Component | Technology |
|---|---|
| AI Backend | Anthropic API (claude-opus-4-6) |
| Structured Outputs | Pydantic v2 + messages.parse() |
| Server-side Search | web_search_20260209 tool |
| Report Generation | python-docx |
| Testing | pytest (52 tests) |
| Thinking Mode | Adaptive thinking |
- Cover Page: Technology name, date, system branding
- Executive Summary: High-level overview
- Technology Overview: How the technology works
- Maturity Assessment: Adoption lifecycle stage
- Market Landscape: Market size, growth, regional split
- Key Players: Major companies with focus areas (table)
- Use Cases: Real-world applications with impact levels (color-coded table)
- Strengths & Limitations: Two-column comparison table
- Adoption Drivers & Barriers: Two-column comparison table
- Key Trends: Current industry developments
- Future Outlook: 3-5 year predictions
- Risk Factors: Threats and challenges
MIT License. See LICENSE