CrewAI AUGMANITAI Example
A multi-agent CrewAI example demonstrating collaborative terminology research using AUGMANITAI.
Architecture
Three agents work sequentially to answer research questions:
- Research Topic Specialist — Analyzes the user’s question and identifies key concepts
- Terminology Researcher — Searches AUGMANITAI for relevant definitions and examples
- Academic Writer — Synthesizes findings into a structured response
Agent Roles
Each agent has distinct role, goal, and backstory that guide its behavior:
- Research Agent: Breaks down complex questions into constituent concepts
- Terminology Agent: Searches and retrieves specialized terminology from AUGMANITAI
- Writer Agent: Synthesizes findings into clear, well-organized responses
Process
The crew uses Process.sequential, meaning:
- Research Agent completes its task first
- Output feeds to Terminology Agent
- Output feeds to Writer Agent
- Final response is returned
Files
agents.py — Agent definitionstasks.py — Task descriptions with expected outputstools.py — AUGMANITAI lookup tool wrappermock_llm.py — Offline mock LLM for democrew_example.py — Crew orchestration and invocationtest_example.py — Pytest testsrun_example.py — CLI runnerrequirements.txt — Dependencies
Running
Install dependencies:
pip install -r requirements.txt
Run example:
Run tests:
pytest test_example.py -v
Extending
To enhance this example:
- Replace
MockLLM with real LLM (e.g., Claude, GPT-4) in agents.py
- Add parallel processing with
Process.hierarchical for complex workflows
- Extend task descriptions to include domain-specific constraints
- Add human review steps between agents
- Integrate additional tools (e.g., web search, academic databases)
- Create specialized agents for specific domains (e.g., legal terminology)
Key Learning Points
- Multi-agent collaboration: How agents with distinct roles can work together
- Sequential processing: Task dependencies and output chaining
- Tool integration: Wrapping external functionality (AUGMANITAI) for agent use
- Task definition: Clear expected outputs guide agent behavior
- Mock LLM patterns: Offline execution for testing and demos
Bound by the Ethical Disclaimer §1–§20 (DISCLAIMER.md) and the repository licenses (LICENSE). Contact and provider: IMPRESSUM.md.