Skip to content

agents.md

Latest commit

 

History

History
74 lines (48 loc) · 2.13 KB

agents.md

File metadata and controls

74 lines (48 loc) · 2.13 KB

Agents

← Back to Main README

Overview

This template includes AI agent functionality with both Streamlit GUI and FastAPI backend options. The agents are built using a modular architecture with support for different agent engines, workflows, and tools.

⚙️ Configuration

First, install the project dependencies:

uv sync

Then, create a .env file in the project root with your API key:

# Create .env file
echo "API_KEY=your_api_key_here" > .env

Replace your_api_key_here with your actual API key for the LLM provider you're using.

🚀 Running the Application

Option 1: Streamlit GUI

To run the application with the Streamlit interface:

uv run streamlit run app/gui.py

This will start the Streamlit development server, typically on http://localhost:8501.

Option 2: FastAPI Backend

To run the FastAPI backend server:

uv run fastapi dev app/main.py --port 8300

This will start the FastAPI development server on port 8300 (avoiding the default 8000 port).

🏗️ Agent Architecture

The agent system is organized into several key components:

Core Components

  • Agent Manager (app/agent/agent_manager.py) - Central management of agent instances
  • Engines (app/agent/engines/) - Different agent execution engines
    • react_agent.py - ReAct (Reasoning and Acting) agent implementation
    • guardrails.py - Safety and validation layer
    • routers.py - Request routing for agent modes
    • translators.py - Input/output translation layer
  • Workflows (app/agent/workflows/) - Agent workflow definitions
  • Tools (app/agent/tools/) - Available tools and utilities
  • Prompts (app/agent/prompts/) - Agent and worker prompt templates

Key Features

  • Modular Design: Easy to extend with new agent types and capabilities
  • Tool Integration: Built-in tool registry and management system
  • Workflow Support: Configurable agent workflows for different use cases
  • Safety Layer: Guardrails for safe agent operation
  • API Integration: RESTful API endpoints for agent interaction

← Back to Main README