Agent Engine 🚀

Agent Engine 🚀


aiagentic-workflowworkflowagents

Agent Engine 🚀

A Python framework for building dynamic, AI-driven agentic workflows with LLM integration

Python Version License: MIT Documentation

Overview

Agent Engine provides a flexible framework for creating AI-powered workflows that dynamically adapt to task requirements. Design complex agent systems with:

  • Customizable agents with specialized tools
  • Dynamic task routing based on workflow state
  • Multi-LLM provider integration
  • Automatic task generation
  • State and memory management

🌟 Why Agent Engine?

FeatureBenefit
🛠️ Dynamic Task RoutingWorkflows adapt to changing requirements
🤖 Multi-LLM OrchestrationMix models from different providers
🧠 Stateful MemoryMaintain context across long workflows
🎨 Visual Workflow DesignConfigure through YAML/JSON or Python API

Ideal for:

✅ Autonomous research assistants

✅ Adaptive customer support systems

✅ Context-aware data processing pipelines

Key Features

Dynamic Workflows

  • Runtime task generation based on objectives
  • Graph-based routing strategies
  • Customizable stopping criteria

🤖 AI Agent System

  • Modular agent/tool architecture
  • Support for multiple LLM providers
  • Context-aware memory management
  • Pre-built templates for common patterns

🔧 Extensible Design

  • Custom agent implementations
  • Plugin system for tools/routing strategies
  • Configuration-driven workflow setup
  • Comprehensive logging and monitoring

Installation

Package is not published yet.

[WIP] pip install agent-engine

⚡ Get Started in 60 Seconds

Create your first AI workflow:

AI Characters Chatting

Project Structure

agent-engine/

├── src/
│   └── agent_engine/
│       ├── __init__.py
│       ├── core/
│       │   ├── __init__.py
│       │   ├── interfaces.py
│       │   ├── models.py
│       │   └── exceptions.py
│       │
│       ├── agents/
│       │   ├── __init__.py
│       │   └── custom/            # Custom agent implementations
│       │
│       ├── tools/
│       │   ├── __init__.py
│       │   ├── console.py         # Console-related tools
│       │   ├── factory.py         # Tool factory
│       │   └── custom/            # Custom tool implementation
│       │
│       ├── routing/
│       │   ├── __init__.py
│       │   ├── graph.py           # Graph-based routing implementation
│       │   └── factory.py         # Routing strategy factory
│       │
│       ├── stopping_criteria/
│       │   ├── __init__.py
│       │   ├── custom.py          # Custom stopping criteria implementation
│       │   └── factory.py         # Stopping criteria factory
│       │
│       ├── config/
│       │   ├── __init__.py
│       │   ├── loader.py          # Configuration loading
│       │   └── models.py          # Configuration schemas
│       │
│       ├── engine/
│       │   ├── __init__.py
│       │   └── workflow_engine.py # Workflow engine implementation
│       │
│       └── utils/
│           ├── __init__.py
│           └── logging.py         # Logging utilities

├── tests/
│   ├── __init__.py
│   ├── conftest.py
│   ├── test_agents/
│   ├── test_tools/
│   ├── test_routing/
│   ├── test_stopping_criteria/
│   └── test_engine/

├── examples/
│   ├── basic_workflow/
│   │   ├── config.yaml
│   │   └── run.py
│   └── topic_research_workflow/
│       ├── config.yaml
│       └── run.py

├── docs/
│   ├── README.md
│   ├── CONTRIBUTING.md
│   └── guides/

├── pyproject.toml
├── setup.py
├── setup.cfg
├── requirements.txt
└── README.md

Contributing

We welcome contributions! Please see:

  • CONTRIBUTING.md for guidelines
  • API Documentation for development details

To add custom components:

  1. Create agents in agents/custom/

  2. Implement tools in tools/custom/

  3. Extend routing strategies in routing/

Documentation

Full documentation available at: 📚 Documentation Portal

© 2025 Aayush Shah