Spaces:
Sleeping
Sleeping
File size: 2,756 Bytes
617ddac 4dbe519 f869dc9 4dbe519 c06d66f 4dbe519 f869dc9 4dbe519 f869dc9 4dbe519 f869dc9 4dbe519 e184879 4dbe519 e184879 4dbe519 f869dc9 4dbe519 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 | ---
title: Agent From Scratch
emoji: 🤖
colorFrom: green
colorTo: blue
sdk: docker
app_port: 7860
---
# AI Agent Framework
A flexible framework for building AI agents with tool support, MCP integration, and multi-step reasoning.
## Features
- **Agent System**: Multi-step reasoning with tool execution
- **Tool Framework**: Easy tool creation and integration
- **MCP Integration**: Load tools from MCP servers
- **LLM Client**: Unified interface for LLM API calls via LiteLLM
- **Modular Design**: Clean, organized package structure
## Installation
### Prerequisites
Install `uv` (recommended package manager):
- macOS/Linux:
```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```
- Homebrew (macOS):
```bash
brew install uv
```
### Setup
1. Create a virtual environment:
```bash
uv venv
source .venv/bin/activate
```
2. Install dependencies:
```bash
uv pip install -r requirements.txt
```
3. Set up environment variables:
```bash
cp .env.example .env
# Edit .env and add your API keys (OPENAI_API_KEY, TAVILY_API_KEY, etc.)
```
## Quick Start
```python
from agent_framework import Agent, LlmClient, FunctionTool
# Define a tool
def calculator(expression: str) -> float:
"""Calculate mathematical expressions."""
return eval(expression)
# Create the agent
agent = Agent(
model=LlmClient(model="gpt-5-mini"),
tools=[FunctionTool(calculator)],
instructions="You are a helpful assistant.",
)
# Run the agent
result = await agent.run("What is 1234 * 5678?")
print(result.output) # "7006652"
```
## Package Structure
```
agent_framework/
├── __init__.py # Package exports
├── models.py # Core data models (Message, ToolCall, Event, ExecutionContext)
├── tools.py # BaseTool and FunctionTool classes
├── llm.py # LlmClient and request/response models
├── agent.py # Agent and AgentResult classes
├── mcp.py # MCP tool loading utilities
└── utils.py # Helper functions for tool definitions
```
## Usage Examples
### Using the @tool Decorator
```python
from agent_framework import tool
@tool
def multiply(a: float, b: float) -> float:
"""Multiply two numbers."""
return a * b
# multiply is now a FunctionTool instance
```
### MCP Tool Integration
```python
from agent_framework import load_mcp_tools
import os
connection = {
"command": "npx",
"args": ["-y", "tavily-mcp@latest"],
"env": {"TAVILY_API_KEY": os.getenv("TAVILY_API_KEY")}
}
mcp_tools = await load_mcp_tools(connection)
agent = Agent(
model=LlmClient(model="gpt-5-mini"),
tools=mcp_tools,
)
```
## Documentation
See `agent_framework/README.md` for detailed API documentation.
## License
See LICENSE file for details.
|