Collection of example agents built with the Agno framework — from simple single-agent scripts to complex multi-agent teams.
| # | File | Type | Description |
|---|---|---|---|
| 1 | src/web_search_agent.py |
Single agent | Web search with DuckDuckGo |
| 2 | src/multi_agents.py |
Multi-agent team | Web + Finance agent team |
| 3 | src/hackernews_analyst.py |
Structured output | HN trend analysis with Pydantic response model |
| 4 | src/research_team.py |
3-agent team | ArXiv + Wikipedia + Web literature review |
| 5 | src/reasoning_agent.py |
Reasoning agent | Chain-of-thought math solver with Calculator + Python |
| 6 | src/data_analyst.py |
Code execution | Python code-writing analyst that runs its own code |
| 7 | src/news_digest.py |
2-agent team | Deep news analysis with article reading (Newspaper4k) |
| 8 | src/investment_team.py |
3-agent team | VC-style investment analysis (fundamentals + sentiment + quant) |
| 9 | src/knowledge_agent.py |
Session memory | Wikipedia Q&A with persistent JSON storage |
| 10 | src/code_generator.py |
Reasoning + code | Self-verifying code writer that plans, builds, and tests |
| 11 | src/medical_research_team.py |
3-agent team | PubMed + ArXiv + Web biomedical literature review |
| 12 | src/startup_scout.py |
3-agent team | VC-style startup evaluation (HN buzz + market + research) |
| 13 | src/content_writer.py |
Reasoning + tools | Research-backed long-form article generator |
| 14 | src/debate_arena.py |
Adversarial team | Pro vs Con debate with moderator judging |
- Python 3.13+
- A Groq API key (required): https://groq.com/
- macOS terminal (zsh) or any POSIX shell
- Optional but recommended: uv (fast Python package and env manager)
-
Install uv (once):
-
Create the environment and install deps (from repo root):
uv sync- Provide your Groq API key (pick one):
-
Export directly in shell:
export GROQ_API_KEY="your_groq_api_key"
-
Or use a .env file with uv:
cp .env.example .env # edit .env and set GROQ_API_KEY export UV_ENV_FILE=".env"
- Run the examples:
# Hello world
uv run python src/main.py
# Web search agent (DuckDuckGo)
uv run python src/web_search_agent.py
# Multi-agent team (web + finance)
uv run python src/multi_agents.py
# HN trend analyst (structured output)
uv run python src/hackernews_analyst.py
# Academic research team (ArXiv + Wikipedia + Web)
uv run python src/research_team.py
# Reasoning math agent (chain-of-thought)
uv run python src/reasoning_agent.py
# Data analyst (code execution)
uv run python src/data_analyst.py
# News digest (deep article reading)
uv run python src/news_digest.py
# Investment advisory team
uv run python src/investment_team.py
# Knowledge explorer (session memory — interactive)
uv run python src/knowledge_agent.py
# Code generator (self-verifying)
uv run python src/code_generator.py
# Medical research team (PubMed + ArXiv)
uv run python src/medical_research_team.py
# Startup scout (VC-style memo)
uv run python src/startup_scout.py
# Content writer (long-form articles)
uv run python src/content_writer.py
# Debate arena (adversarial debate)
uv run python src/debate_arena.pyIf you prefer not to use uv:
# Create and activate a venv
python3 -m venv .venv
source .venv/bin/activate
# Install the package (and its dependencies)
pip install .
# Set your Groq key
export GROQ_API_KEY="your_groq_api_key"
# Run examples
python src/web_search_agent.py
python src/multi_agents.py