AgenticAI-App

This repository contains the application layer built on top of the AgenticAI framework. While the AgenticAI framework provides the core abstractions for agents, tools, planning, memory, and workflow orchestration, this repository demonstrates how to build a runnable application using those components.

The app includes a simple command-line interface (CLI) that uses the ResearchAgent and built-in tools from the AgenticAI framework. It also includes a suite of workflow tests that validate multi-step reasoning, refinement, dependencies, and conditional execution.

---

Purpose of This Repository

• Provide a runnable example of how to use the AgenticAI framework
• Demonstrate how to assemble agents, tools, and workflows into an application
• Serve as a testing ground for workflow logic and multi-step reasoning
• Offer a clean starting point for building more advanced agentic applications

This repository depends on the AgenticAI framework, which must be installed separately.

---

Project Structure

src/
    agenticai_app/
        main.py            – CLI entry point for running the ResearchAgent
        __init__.py        – Package initializer

    cli/                   – Placeholder for future CLI commands
    configs/               – Placeholder for configuration files
    workflows/             – Placeholder for workflow definitions

tests/
    test_step_2_4.py
    test_workflow_agent_basic.py
    test_workflow_agent_conditionals.py
    test_workflow_agent_dependencies.py
    test_workflow_agent_dependencies_out_of_order.py
    test_workflow_agent_final_synthesis.py
    test_workflow_agent_global_refinement.py
    test_workflow_agent_parallel.py
    test_workflow_agent_refinement.py
    test_workflow_agent_while.py

docs/
    Notes.txt              – Personal notes (ignored by Git)

pyproject.toml             – Package metadata  
requirements.txt           – Python dependencies  
README.md                  – This file  
.gitignore                 – Ignore rules  

---

Installation

Before running the app, install the AgenticAI framework. If you have cloned the framework repository locally, install it in editable mode:

pip install -e ../AgenticAI

Or install directly from GitHub once the framework is published:

pip install git+https://github.com/YOUR_USERNAME/AgenticAI.git

Then install the app dependencies:

pip install -r requirements.txt

---

Running the Application

To start the interactive ResearchAgent CLI:

python -m agenticai_app.main

You will see:

ResearchAgent ready. Type 'exit' to quit.

You can then type natural language queries, and the agent will respond using tools or direct reasoning.

---

Example Interaction

You: define machine learning
[Tools used: DefinitionTool] Machine learning is...

You: summarize the history of quantum computing
[Tools used: SummarizerTool] Quantum computing began...

You: compare supervised and unsupervised learning
It looks like you're asking for a multi-step task.

Raw steps:
1. compare supervised learning
2. compare unsupervised learning

Parsed steps:
1. action=compare, target=supervised learning, tool=None
2. action=compare, target=unsupervised learning, tool=None

---

Running Tests

This repository includes a suite of workflow tests that validate:

• step parsing
• refinement loops
• dependency ordering
• conditional execution
• parallel execution
• final synthesis

To run all tests:

pytest

---

Design Philosophy

The app follows the same principles as the AgenticAI framework:

1. Simplicity

The application should be easy to run and modify.

2. Modularity

CLI, workflows, and configuration are separated cleanly.

3. Extensibility

New workflows, tools, and agents can be added without modifying core logic.

4. Transparency

The app exposes intermediate reasoning steps when appropriate.

---

Roadmap

Iteration 2

• Add real CLI commands under src/cli/
• Add configuration loading under src/configs/
• Add workflow templates under src/workflows/
• Add integration tests for CLI behavior

Iteration 3

• Add multi-agent orchestration demos
• Add advanced workflow execution examples
• Add logging and tracing utilities

Iteration 4+

• Add GUI or web-based interface
• Add plugin system for custom workflows
• Add deployment templates

---

License

This project is licensed under the MIT License.