Examples
The examples/
directory contains twelve runnable scripts that build up from a one-line agent
to autonomous, self-improving systems. Each file is self-contained and
documents its own install and API-key requirements in the docstring at the top.
Running an Example
Some examples need extra dependencies or keys (noted below and in each file) —
for instance web search needs tavily-python and a TAVILY_API_KEY.
The Scripts
Fundamentals
| # | Script | What it shows |
|---|---|---|
| 01 | 01_basic_agent.py |
The minimal agent — a model plus a prompt. See Agents. |
| 02 | 02_agent_with_tools.py |
Attaching built-in tools (web search + Python execution). See Tools. |
| 03 | 03_agent_with_memory.py |
Persistent memory so the agent recalls earlier turns. See Memory. |
| 10 | 10_custom_tool.py |
Turning a plain Python function into an agent tool. See Tools. |
Models & Output
| # | Script | What it shows |
|---|---|---|
| 08 | 08_multi_model_comparison.py |
Running one prompt across OpenAI, Anthropic, and Google to compare results. See Models. |
| 09 | 09_streaming_agent.py |
Real-time token streaming with run(..., stream=True) — ideal for chat UIs. |
Knowledge & Reasoning
| # | Script | What it shows |
|---|---|---|
| 05 | 05_rag_with_irag.py |
Retrieval-Augmented Generation with the built-in iRAG engine (no vector DB). See Knowledge. |
| 06 | 06_planning_agent.py |
PlanningAgent decomposing a goal into steps, then executing and monitoring them. |
Multi-Agent & Deployment
| # | Script | What it shows |
|---|---|---|
| 04 | 04_multi_agent_team.py |
A Team where a coordinator delegates to specialist agents. See Models & Teams. |
| 07 | 07_fastapi_deployment.py |
Serving an agent as a FastAPI REST API. See Quick Start. |
Autonomous Systems
| # | Script | What it shows |
|---|---|---|
| 11 | 11_pulse_employee.py |
The PULSE virtual employee: KT sessions, meetings, tasks, and status updates. See PULSE. |
| 12 | 12_competency_engine.py |
Scoring competency, deficit-driven learning, and runtime routing. See Competency Engine. |
Start small
If you're new, run 01 → 02 → 03 in order. They introduce the model,
tools, and memory concepts that every later example builds on.