Define tools
Describe Python functions with the tool decorator and keep execution explicit.
Small pieces for tool-using agents
An agent loop you can read.
A Python framework for model calls, typed tools, conversation memory and middleware, with an OpenAI-compatible client.
pyagentPROJECT OVERVIEW
01Define a tool
02Run the agent
03Inspect the result
Illustrated workflow · not a live application session
01 / Inside the project
Built around the work.
Start with the implemented workflow, then follow the source to understand how it works.
Connect a model
Use an OpenAI-compatible endpoint and configure the provider yourself.
Inspect the loop
Agent, parsing, memory and middleware live in separate, readable modules.
02 / Explore the workflow
Three steps to understand it.
Define tools
Describe Python functions with the tool decorator and keep execution explicit.
Full setup and implementation ↗03 / Run it yourself
Start from the source.
See pyproject.toml for supported Python versions. The example requires an API key or a configured local endpoint.
Clone the repository and run these commands from its root. Review the README before connecting external services.
Open repository ↗SOURCE CHECKOUT
git clone https://github.com/honeyamn10-source/pyagent.git
cd pyagent
python -m pip install .
python examples/calculator_agent.py04 / Know the scope
Clear expectations.
Model quality and costs depend on the provider. Tools run with the permissions you give their Python process; this library is not an operating-system sandbox.
Check the latest results in GitHub Actions. A passing automated check covers its tested paths, not every real-world integration.