Metadata-Version: 2.4
Name: agentflowpy
Version: 0.1.2
Summary: Practial module to build AI agents with context management and step-based execution.
Author-email: Ozel Umut Tam <ozel.umut.tam@gmail.com>
License: NonCommercial No-AI License
Requires-Python: >=3.12
Description-Content-Type: text/markdown
License-File: LICENSE.md
Requires-Dist: pydantic>=2.11.0
Provides-Extra: dev
Requires-Dist: pytest>=7.0.0; extra == "dev"
Dynamic: license-file

# AgentFlowPy

**AgentFlowPy** is a lightweight Python framework for building step-based workflows with flexible context management. It allows developers to create agents that execute sequences of steps, maintain context, pass arguments between steps, and support typed messages using Pydantic models.

## Features

- Step-based workflow execution.
- Flexible **context management** with multiple contexts.
- Pass arguments between steps using `StepPass`.
- Typed messages with Pydantic support.
- Easy integration of custom logic in each step.
- Looping and conditional flow control (`AGENT_START`, `AGENT_END`).

## Installation

```bash
pip install agentflowpy
```

## Examples
- **Basic Usage :** [basic_flow.py](./examples/basic_flow.py) *Sample of basic use.*
- **Typed Message :** [typed_message.py](./examples/typed_message.py) *Typed Message and Conditional flow.*
- **Pass Arguments :** [pass_args.py](./examples/pass_args.py) *Passing additional arguments to next step.*
- **Context Switch In Flow :** [mid_switch.py](./examples/mid_switch.py) *Switch context while in flow to effect next steps.*
- **Serialize/Restore Contexts:** [save_load_state.py](./examples/save_load_state.py) *Serialize and Restore context_manager state*
