| Name | T-AI-project JSON |
| Version |
0.1.0
JSON |
| download |
| home_page | None |
| Summary | A Simplified Agent Framework to create ai agent with tools,mcps and memory/deps management |
| upload_time | 2025-07-28 16:27:51 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.9 |
| license | MIT |
| keywords |
ai
assistant
google-ai
mcp
openai
pydantic-ai
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# T_AI - AI Agent Framework
A powerful framework for building AI agents with **MCP (Model Context Protocol)** integration, **tools**, **memory management**, and **dependency handling**. T_AI simplifies the creation of sophisticated AI agents by providing a unified interface for multiple LLMs, external tools, and advanced conversation management.
## 🚀 Key Features
- **🔗 MCP Integration**: Seamless connection to external tools and services via Model Context Protocol
- **🛠️ Tool System**: Built-in Google tools (image search, code execution) and easy custom tool integration
- **🧠 Memory Management**: Intelligent conversation summarization for long-running sessions
- **📦 Dependency Management**: Clean state management and user context handling
- **🤖 Multi-LLM Support**: Compatible with Google Gemini, OpenAI, and Anthropic models
- **📱 Media Support**: Handle text, audio, images, and PDF files seamlessly
- **⚡ Async/Await**: Full asynchronous support for optimal performance
- **🔌 Extensible**: Easy to extend with custom tools and integrations
## 📦 Installation
### Using UV (Recommended)
```bash
# Clone the repository
git clone <repository-url>
cd T_AI
# Install using UV
uv sync
```
### Using pip
```bash
pip install -e .
```
### Using pip with requirements.txt
```bash
pip install -r requirements.txt
```
## 🔧 Core Dependencies
- **pydantic-ai >= 0.4.0**: Core AI framework
- **tavily-python >= 0.5.1**: Web search capabilities
## 🚀 Quick Start
### Basic Agent Creation
```python
import asyncio
from t_ai.t_agent import TAgent, Deps
from pydantic_ai.models.google import GoogleModel
from pydantic_ai.providers.google import GoogleProvider
# Initialize with Google Gemini
llm = GoogleModel('gemini-2.5-flash', provider=GoogleProvider(api_key="your-api-key"))
agent = TAgent(llm=llm)
# Simple conversation
async def main():
async with agent:
response = await agent.chat(["Hello, what can you help me with?"])
print(f"UI Version: {response.ui_version}")
print(f"Voice Version: {response.voice_version}")
asyncio.run(main())
```
### With OpenAI
```python
from pydantic_ai.models.openai import OpenAIModel
from pydantic_ai.providers.openai import OpenAIProvider
llm = OpenAIModel('gpt-4-mini', provider=OpenAIProvider(api_key="your-openai-key"))
agent = TAgent(llm=llm)
```
## 🛠️ Framework Configuration
### TAgent Parameters
```python
agent = TAgent(
llm=your_llm, # Required: pydantic-ai compatible model
deps=custom_deps, # Optional: Custom dependencies object
instructions="Custom instructions", # Optional: Agent instructions
tools=[], # Optional: List of custom tools
mcp_servers=[], # Optional: List of MCP servers
summarizer=False, # Optional: Enable conversation summarization
custom_summarizer_agent=None, # Optional: Custom summarizer agent
memory_length=20, # Optional: Messages before summarization
memory_summarizer_length=15, # Optional: Messages to summarize
use_memory=True # Optional: Enable/disable memory
)
```
## 🔗 MCP Server Integration
### Using MCP Helper (Recommended)
```python
from t_ai.utils.helper_functions import MCP_server_helper
from t_ai.t_agent import TAgent
# Create MCP helper
mcp_helper = MCP_server_helper()
# Add different types of MCP servers
mcp_helper.add_mpc_server(type='http', mpc_server_url='https://mcp.notion.com/mcp')
mcp_helper.add_mpc_server(type='sse', mpc_server_url='https://mcp.notion.com/sse')
mcp_helper.add_mpc_server(type='stdio', command='npx', args=['-y', 'mcp-remote', 'https://mcp.notion.com/mcp'])
# Initialize agent with MCP servers
agent = TAgent(llm=llm, mcp_servers=mcp_helper.get_mpc_servers())
```
### Direct MCP Server Setup
```python
from pydantic_ai.mcp import MCPServerStreamableHTTP, MCPServerSSE, MCPServerStdio
mcp_servers = [
MCPServerStreamableHTTP(url='https://mcp.notion.com/mcp'),
MCPServerSSE(url='https://mcp.notion.com/sse'),
MCPServerStdio(command='npx', args=['-y', 'mcp-remote', 'https://mcp.notion.com/mcp'])
]
agent = TAgent(llm=llm, mcp_servers=mcp_servers)
```
## 🛠️ Built-in Tools
### Google Image Search Tool
```python
from t_ai.PrebuiltTools.google_tools import search_images_tool
# Setup image search
image_tool = search_images_tool(
api_key="your-google-api-key",
search_engine_id="your-custom-search-engine-id"
)
agent = TAgent(llm=llm, tools=[image_tool])
# Usage
response = await agent.chat(["Find me an image of a sunset"])
```
### Google Code Execution Tool
```python
from t_ai.PrebuiltTools.google_tools import code_execution_tool
# Setup code execution
code_tool = code_execution_tool(api_key="your-gemini-api-key")
agent = TAgent(llm=llm, tools=[code_tool])
# Usage
response = await agent.chat(["Calculate the factorial of 10 using Python"])
```
### Combined Tools Example
```python
from t_ai.PrebuiltTools.google_tools import search_images_tool, code_execution_tool
tools = [
search_images_tool(api_key=google_api_key, search_engine_id=search_engine_id),
code_execution_tool(api_key=google_api_key)
]
agent = TAgent(llm=llm, tools=tools)
```
## 💾 Memory Management
### Enable Automatic Summarization
```python
agent = TAgent(
llm=llm,
summarizer=True, # Enable summarization
memory_length=20, # Summarize after 20 messages
memory_summarizer_length=15 # Summarize oldest 15 messages
)
```
### Custom Summarizer Agent
```python
from pydantic_ai import Agent
custom_summarizer = Agent(
llm,
instructions='Create detailed technical summaries focusing on code and solutions.'
)
agent = TAgent(
llm=llm,
summarizer=True,
custom_summarizer_agent=custom_summarizer
)
```
### Memory and State Management
```python
# Access conversation history
messages = agent.memory.messages
# Access agent dependencies
deps = agent.deps
user_name = agent.deps.user
agents_output = agent.deps.agents_output
# Reset agent state
agent.reset()
```
## 📱 Media Support
### Text Input
```python
response = await agent.chat(["What's the weather like today?"])
```
### Image Input
```python
from pydantic_ai.messages import BinaryContent
# From file
with open("image.png", "rb") as f:
image_data = f.read()
response = await agent.chat([
"What do you see in this image?",
BinaryContent(data=image_data, media_type='image/png')
])
```
### Audio Input
```python
# Audio file
with open("audio.wav", "rb") as f:
audio_data = f.read()
response = await agent.chat([
"Transcribe this audio",
BinaryContent(data=audio_data, media_type='audio/wav')
])
```
### PDF Input
```python
# PDF file
with open("document.pdf", "rb") as f:
pdf_data = f.read()
response = await agent.chat([
"Summarize this document",
BinaryContent(data=pdf_data, media_type='application/pdf')
])
```
## 🔧 Advanced Usage
### Context Manager (Recommended)
```python
async def main():
async with TAgent(llm=llm, mcp_servers=mcp_servers) as agent:
# MCP servers are automatically connected
response = await agent.chat(["Help me with my Notion workspace"])
print(response.ui_version)
# MCP servers are automatically disconnected
```
### Manual Connection Management
```python
agent = TAgent(llm=llm, mcp_servers=mcp_servers)
# Connect manually
await agent.connect()
try:
response = await agent.chat(["Hello"])
finally:
# Disconnect manually
await agent.disconnect()
```
### Custom Dependencies
```python
from t_ai.t_agent import Deps
# Create custom dependencies
custom_deps = Deps(
agents_output={"previous_results": []},
user="Alice"
)
agent = TAgent(llm=llm, deps=custom_deps)
```
### Custom Tools
```python
from pydantic_ai.tools import Tool
def custom_weather_tool(location: str) -> str:
"""Get weather information for a location"""
# Your weather API logic here
return f"Weather in {location}: Sunny, 25°C"
weather_tool = Tool(
custom_weather_tool,
name='get_weather',
description='Get current weather for any location'
)
agent = TAgent(llm=llm, tools=[weather_tool])
```
## 📝 Complete Framework Example
```python
import asyncio
import os
from dotenv import load_dotenv
from t_ai.t_agent import TAgent, Deps
from t_ai.utils.helper_functions import MCP_server_helper
from t_ai.PrebuiltTools.google_tools import search_images_tool, code_execution_tool
from pydantic_ai.models.google import GoogleModel
from pydantic_ai.providers.google import GoogleProvider
from pydantic_ai.messages import BinaryContent
# Load environment variables
load_dotenv()
async def main():
# Setup LLM
llm = GoogleModel('gemini-2.5-flash',
provider=GoogleProvider(api_key=os.getenv('GOOGLE_API_KEY')))
# Setup MCP servers
mcp_helper = MCP_server_helper()
mcp_helper.add_mpc_server(type='stdio', command='npx',
args=['-y', '@modelcontextprotocol/server-filesystem', '/tmp'])
# Setup tools
tools = [
search_images_tool(
api_key=os.getenv('GOOGLE_API_KEY'),
search_engine_id=os.getenv('GOOGLE_SEARCH_ENGINE_ID')
),
code_execution_tool(api_key=os.getenv('GOOGLE_API_KEY'))
]
# Setup custom dependencies
deps = Deps(agents_output={}, user="Alice")
# Initialize T_AI agent
agent = TAgent(
llm=llm,
deps=deps,
tools=tools,
mcp_servers=mcp_helper.get_mpc_servers(),
summarizer=True,
memory_length=20,
instructions="You are a helpful AI assistant with access to various tools and services."
)
# Use context manager for automatic connection handling
async with agent:
# Text conversation
response = await agent.chat(["Hello, what can you help me with?"])
print("Agent:", response.voice_version)
# Math problem with code execution
response = await agent.chat(["Calculate the sum of squares from 1 to 100"])
print("Math Result:", response.ui_version)
# Image search
response = await agent.chat(["Find me an image of a beautiful landscape"])
print("Image Search:", response.ui_version)
# Check conversation history
print(f"Total messages in memory: {len(agent.memory.messages)}")
if __name__ == "__main__":
asyncio.run(main())
```
## 🧪 Testing and Development
Run the included Jupyter notebooks to test different features:
- `notebooks/cortana_test.ipynb`: Basic functionality testing
- `notebooks/cort_mcp_test.ipynb`: MCP server integration testing
- `notebooks/cortana_voice_test.ipynb`: Voice/audio capabilities testing
- `notebooks/memory_handling.ipynb`: Memory management testing
## 🔑 Environment Variables
Create a `.env` file in your project root:
```env
GOOGLE_API_KEY=your_google_api_key
GOOGLE_SEARCH_ENGINE_ID=your_custom_search_engine_id
OPENAI_API_KEY=your_openai_api_key
TAVILY_API_KEY=your_tavily_api_key
```
## 🏗️ Architecture
T_AI is built with a modular architecture:
- **Core Agent (`TAgent`)**: Main framework class handling LLM interactions, memory, and coordination
- **MCP Integration**: Support for Model Context Protocol servers (HTTP, SSE, stdio)
- **Tool System**: Extensible tool framework with built-in Google tools
- **Memory Management**: Intelligent conversation summarization and state management
- **Dependencies**: Clean dependency injection for user context and shared state
## 🤝 Contributing
1. Fork the repository
2. Create a feature branch (`git checkout -b feature/amazing-feature`)
3. Make your changes
4. Add tests if applicable
5. Commit your changes (`git commit -m 'Add amazing feature'`)
6. Push to the branch (`git push origin feature/amazing-feature`)
7. Open a Pull Request
## 📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
## 🆘 Support
For issues and questions:
1. Check the notebooks in the `notebooks/` directory for examples
2. Review the docstrings in the source code
3. Open an issue on GitHub
## 🙏 Acknowledgments
- Built on top of [pydantic-ai](https://github.com/pydantic/pydantic-ai)
- MCP (Model Context Protocol) integration
- Google AI and OpenAI API support
Raw data
{
"_id": null,
"home_page": null,
"name": "T-AI-project",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "ai, assistant, google-ai, mcp, openai, pydantic-ai",
"author": null,
"author_email": "Tristan Padiou <padioutristan@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/1c/71/ef9f36ee7e26adb00abaa6d2f576f8e61b5e9863f60604ebfc41b996073f/t_ai_project-0.1.0.tar.gz",
"platform": null,
"description": "# T_AI - AI Agent Framework\n\nA powerful framework for building AI agents with **MCP (Model Context Protocol)** integration, **tools**, **memory management**, and **dependency handling**. T_AI simplifies the creation of sophisticated AI agents by providing a unified interface for multiple LLMs, external tools, and advanced conversation management.\n\n## \ud83d\ude80 Key Features\n\n- **\ud83d\udd17 MCP Integration**: Seamless connection to external tools and services via Model Context Protocol\n- **\ud83d\udee0\ufe0f Tool System**: Built-in Google tools (image search, code execution) and easy custom tool integration \n- **\ud83e\udde0 Memory Management**: Intelligent conversation summarization for long-running sessions\n- **\ud83d\udce6 Dependency Management**: Clean state management and user context handling\n- **\ud83e\udd16 Multi-LLM Support**: Compatible with Google Gemini, OpenAI, and Anthropic models\n- **\ud83d\udcf1 Media Support**: Handle text, audio, images, and PDF files seamlessly\n- **\u26a1 Async/Await**: Full asynchronous support for optimal performance\n- **\ud83d\udd0c Extensible**: Easy to extend with custom tools and integrations\n\n## \ud83d\udce6 Installation\n\n### Using UV (Recommended)\n\n```bash\n# Clone the repository\ngit clone <repository-url>\ncd T_AI\n\n# Install using UV\nuv sync\n```\n\n### Using pip\n\n```bash\npip install -e .\n```\n\n### Using pip with requirements.txt\n\n```bash\npip install -r requirements.txt\n```\n\n## \ud83d\udd27 Core Dependencies\n\n- **pydantic-ai >= 0.4.0**: Core AI framework\n- **tavily-python >= 0.5.1**: Web search capabilities\n\n## \ud83d\ude80 Quick Start\n\n### Basic Agent Creation\n\n```python\nimport asyncio\nfrom t_ai.t_agent import TAgent, Deps\nfrom pydantic_ai.models.google import GoogleModel\nfrom pydantic_ai.providers.google import GoogleProvider\n\n# Initialize with Google Gemini\nllm = GoogleModel('gemini-2.5-flash', provider=GoogleProvider(api_key=\"your-api-key\"))\nagent = TAgent(llm=llm)\n\n# Simple conversation\nasync def main():\n async with agent:\n response = await agent.chat([\"Hello, what can you help me with?\"])\n print(f\"UI Version: {response.ui_version}\")\n print(f\"Voice Version: {response.voice_version}\")\n\nasyncio.run(main())\n```\n\n### With OpenAI\n\n```python\nfrom pydantic_ai.models.openai import OpenAIModel\nfrom pydantic_ai.providers.openai import OpenAIProvider\n\nllm = OpenAIModel('gpt-4-mini', provider=OpenAIProvider(api_key=\"your-openai-key\"))\nagent = TAgent(llm=llm)\n```\n\n## \ud83d\udee0\ufe0f Framework Configuration\n\n### TAgent Parameters\n\n```python\nagent = TAgent(\n llm=your_llm, # Required: pydantic-ai compatible model\n deps=custom_deps, # Optional: Custom dependencies object\n instructions=\"Custom instructions\", # Optional: Agent instructions\n tools=[], # Optional: List of custom tools\n mcp_servers=[], # Optional: List of MCP servers\n summarizer=False, # Optional: Enable conversation summarization\n custom_summarizer_agent=None, # Optional: Custom summarizer agent\n memory_length=20, # Optional: Messages before summarization\n memory_summarizer_length=15, # Optional: Messages to summarize\n use_memory=True # Optional: Enable/disable memory\n)\n```\n\n## \ud83d\udd17 MCP Server Integration\n\n### Using MCP Helper (Recommended)\n\n```python\nfrom t_ai.utils.helper_functions import MCP_server_helper\nfrom t_ai.t_agent import TAgent\n\n# Create MCP helper\nmcp_helper = MCP_server_helper()\n\n# Add different types of MCP servers\nmcp_helper.add_mpc_server(type='http', mpc_server_url='https://mcp.notion.com/mcp')\nmcp_helper.add_mpc_server(type='sse', mpc_server_url='https://mcp.notion.com/sse')\nmcp_helper.add_mpc_server(type='stdio', command='npx', args=['-y', 'mcp-remote', 'https://mcp.notion.com/mcp'])\n\n# Initialize agent with MCP servers\nagent = TAgent(llm=llm, mcp_servers=mcp_helper.get_mpc_servers())\n```\n\n### Direct MCP Server Setup\n\n```python\nfrom pydantic_ai.mcp import MCPServerStreamableHTTP, MCPServerSSE, MCPServerStdio\n\nmcp_servers = [\n MCPServerStreamableHTTP(url='https://mcp.notion.com/mcp'),\n MCPServerSSE(url='https://mcp.notion.com/sse'),\n MCPServerStdio(command='npx', args=['-y', 'mcp-remote', 'https://mcp.notion.com/mcp'])\n]\n\nagent = TAgent(llm=llm, mcp_servers=mcp_servers)\n```\n\n## \ud83d\udee0\ufe0f Built-in Tools\n\n### Google Image Search Tool\n\n```python\nfrom t_ai.PrebuiltTools.google_tools import search_images_tool\n\n# Setup image search\nimage_tool = search_images_tool(\n api_key=\"your-google-api-key\",\n search_engine_id=\"your-custom-search-engine-id\"\n)\n\nagent = TAgent(llm=llm, tools=[image_tool])\n\n# Usage\nresponse = await agent.chat([\"Find me an image of a sunset\"])\n```\n\n### Google Code Execution Tool\n\n```python\nfrom t_ai.PrebuiltTools.google_tools import code_execution_tool\n\n# Setup code execution \ncode_tool = code_execution_tool(api_key=\"your-gemini-api-key\")\n\nagent = TAgent(llm=llm, tools=[code_tool])\n\n# Usage\nresponse = await agent.chat([\"Calculate the factorial of 10 using Python\"])\n```\n\n### Combined Tools Example\n\n```python\nfrom t_ai.PrebuiltTools.google_tools import search_images_tool, code_execution_tool\n\ntools = [\n search_images_tool(api_key=google_api_key, search_engine_id=search_engine_id),\n code_execution_tool(api_key=google_api_key)\n]\n\nagent = TAgent(llm=llm, tools=tools)\n```\n\n## \ud83d\udcbe Memory Management\n\n### Enable Automatic Summarization\n\n```python\nagent = TAgent(\n llm=llm,\n summarizer=True, # Enable summarization\n memory_length=20, # Summarize after 20 messages\n memory_summarizer_length=15 # Summarize oldest 15 messages\n)\n```\n\n### Custom Summarizer Agent\n\n```python\nfrom pydantic_ai import Agent\n\ncustom_summarizer = Agent(\n llm, \n instructions='Create detailed technical summaries focusing on code and solutions.'\n)\n\nagent = TAgent(\n llm=llm,\n summarizer=True,\n custom_summarizer_agent=custom_summarizer\n)\n```\n\n### Memory and State Management\n\n```python\n# Access conversation history\nmessages = agent.memory.messages\n\n# Access agent dependencies\ndeps = agent.deps\nuser_name = agent.deps.user\nagents_output = agent.deps.agents_output\n\n# Reset agent state\nagent.reset()\n```\n\n## \ud83d\udcf1 Media Support\n\n### Text Input\n\n```python\nresponse = await agent.chat([\"What's the weather like today?\"])\n```\n\n### Image Input\n\n```python\nfrom pydantic_ai.messages import BinaryContent\n\n# From file\nwith open(\"image.png\", \"rb\") as f:\n image_data = f.read()\n\nresponse = await agent.chat([\n \"What do you see in this image?\",\n BinaryContent(data=image_data, media_type='image/png')\n])\n```\n\n### Audio Input\n\n```python\n# Audio file\nwith open(\"audio.wav\", \"rb\") as f:\n audio_data = f.read()\n\nresponse = await agent.chat([\n \"Transcribe this audio\",\n BinaryContent(data=audio_data, media_type='audio/wav')\n])\n```\n\n### PDF Input\n\n```python\n# PDF file\nwith open(\"document.pdf\", \"rb\") as f:\n pdf_data = f.read()\n\nresponse = await agent.chat([\n \"Summarize this document\",\n BinaryContent(data=pdf_data, media_type='application/pdf')\n])\n```\n\n## \ud83d\udd27 Advanced Usage\n\n### Context Manager (Recommended)\n\n```python\nasync def main():\n async with TAgent(llm=llm, mcp_servers=mcp_servers) as agent:\n # MCP servers are automatically connected\n response = await agent.chat([\"Help me with my Notion workspace\"])\n print(response.ui_version)\n # MCP servers are automatically disconnected\n```\n\n### Manual Connection Management\n\n```python\nagent = TAgent(llm=llm, mcp_servers=mcp_servers)\n\n# Connect manually\nawait agent.connect()\n\ntry:\n response = await agent.chat([\"Hello\"])\nfinally:\n # Disconnect manually\n await agent.disconnect()\n```\n\n### Custom Dependencies\n\n```python\nfrom t_ai.t_agent import Deps\n\n# Create custom dependencies\ncustom_deps = Deps(\n agents_output={\"previous_results\": []},\n user=\"Alice\"\n)\n\nagent = TAgent(llm=llm, deps=custom_deps)\n```\n\n### Custom Tools\n\n```python\nfrom pydantic_ai.tools import Tool\n\ndef custom_weather_tool(location: str) -> str:\n \"\"\"Get weather information for a location\"\"\"\n # Your weather API logic here\n return f\"Weather in {location}: Sunny, 25\u00b0C\"\n\nweather_tool = Tool(\n custom_weather_tool,\n name='get_weather',\n description='Get current weather for any location'\n)\n\nagent = TAgent(llm=llm, tools=[weather_tool])\n```\n\n## \ud83d\udcdd Complete Framework Example\n\n```python\nimport asyncio\nimport os\nfrom dotenv import load_dotenv\n\nfrom t_ai.t_agent import TAgent, Deps\nfrom t_ai.utils.helper_functions import MCP_server_helper\nfrom t_ai.PrebuiltTools.google_tools import search_images_tool, code_execution_tool\n\nfrom pydantic_ai.models.google import GoogleModel\nfrom pydantic_ai.providers.google import GoogleProvider\nfrom pydantic_ai.messages import BinaryContent\n\n# Load environment variables\nload_dotenv()\n\nasync def main():\n # Setup LLM\n llm = GoogleModel('gemini-2.5-flash', \n provider=GoogleProvider(api_key=os.getenv('GOOGLE_API_KEY')))\n \n # Setup MCP servers\n mcp_helper = MCP_server_helper()\n mcp_helper.add_mpc_server(type='stdio', command='npx', \n args=['-y', '@modelcontextprotocol/server-filesystem', '/tmp'])\n \n # Setup tools\n tools = [\n search_images_tool(\n api_key=os.getenv('GOOGLE_API_KEY'),\n search_engine_id=os.getenv('GOOGLE_SEARCH_ENGINE_ID')\n ),\n code_execution_tool(api_key=os.getenv('GOOGLE_API_KEY'))\n ]\n \n # Setup custom dependencies\n deps = Deps(agents_output={}, user=\"Alice\")\n \n # Initialize T_AI agent\n agent = TAgent(\n llm=llm,\n deps=deps,\n tools=tools,\n mcp_servers=mcp_helper.get_mpc_servers(),\n summarizer=True,\n memory_length=20,\n instructions=\"You are a helpful AI assistant with access to various tools and services.\"\n )\n \n # Use context manager for automatic connection handling\n async with agent:\n # Text conversation\n response = await agent.chat([\"Hello, what can you help me with?\"])\n print(\"Agent:\", response.voice_version)\n \n # Math problem with code execution\n response = await agent.chat([\"Calculate the sum of squares from 1 to 100\"])\n print(\"Math Result:\", response.ui_version)\n \n # Image search\n response = await agent.chat([\"Find me an image of a beautiful landscape\"])\n print(\"Image Search:\", response.ui_version)\n \n # Check conversation history\n print(f\"Total messages in memory: {len(agent.memory.messages)}\")\n\nif __name__ == \"__main__\":\n asyncio.run(main())\n```\n\n## \ud83e\uddea Testing and Development\n\nRun the included Jupyter notebooks to test different features:\n\n- `notebooks/cortana_test.ipynb`: Basic functionality testing\n- `notebooks/cort_mcp_test.ipynb`: MCP server integration testing \n- `notebooks/cortana_voice_test.ipynb`: Voice/audio capabilities testing\n- `notebooks/memory_handling.ipynb`: Memory management testing\n\n## \ud83d\udd11 Environment Variables\n\nCreate a `.env` file in your project root:\n\n```env\nGOOGLE_API_KEY=your_google_api_key\nGOOGLE_SEARCH_ENGINE_ID=your_custom_search_engine_id\nOPENAI_API_KEY=your_openai_api_key\nTAVILY_API_KEY=your_tavily_api_key\n```\n\n## \ud83c\udfd7\ufe0f Architecture\n\nT_AI is built with a modular architecture:\n\n- **Core Agent (`TAgent`)**: Main framework class handling LLM interactions, memory, and coordination\n- **MCP Integration**: Support for Model Context Protocol servers (HTTP, SSE, stdio)\n- **Tool System**: Extensible tool framework with built-in Google tools\n- **Memory Management**: Intelligent conversation summarization and state management\n- **Dependencies**: Clean dependency injection for user context and shared state\n\n## \ud83e\udd1d Contributing\n\n1. Fork the repository\n2. Create a feature branch (`git checkout -b feature/amazing-feature`)\n3. Make your changes\n4. Add tests if applicable\n5. Commit your changes (`git commit -m 'Add amazing feature'`)\n6. Push to the branch (`git push origin feature/amazing-feature`)\n7. Open a Pull Request\n\n## \ud83d\udcc4 License\n\nThis project is licensed under the MIT License - see the LICENSE file for details.\n\n## \ud83c\udd98 Support\n\nFor issues and questions:\n1. Check the notebooks in the `notebooks/` directory for examples\n2. Review the docstrings in the source code \n3. Open an issue on GitHub\n\n## \ud83d\ude4f Acknowledgments\n\n- Built on top of [pydantic-ai](https://github.com/pydantic/pydantic-ai)\n- MCP (Model Context Protocol) integration\n- Google AI and OpenAI API support\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A Simplified Agent Framework to create ai agent with tools,mcps and memory/deps management",
"version": "0.1.0",
"project_urls": {
"Bug Tracker": "https://github.com/yourusername/cortana-agent/issues",
"Documentation": "https://github.com/yourusername/cortana-agent#readme",
"Homepage": "https://github.com/yourusername/cortana-agent",
"Repository": "https://github.com/yourusername/cortana-agent"
},
"split_keywords": [
"ai",
" assistant",
" google-ai",
" mcp",
" openai",
" pydantic-ai"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "4b669a215ce40ae7079c0b865a86c299f53cf41f750eb56d8f8282c2fb75208f",
"md5": "b100429ab12c7911bacc11475c52d51f",
"sha256": "6479df4e2b647c60995eb2f9ba68eacd66423db5045a6e23385764aba9e916b5"
},
"downloads": -1,
"filename": "t_ai_project-0.1.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "b100429ab12c7911bacc11475c52d51f",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 11656,
"upload_time": "2025-07-28T16:27:50",
"upload_time_iso_8601": "2025-07-28T16:27:50.767482Z",
"url": "https://files.pythonhosted.org/packages/4b/66/9a215ce40ae7079c0b865a86c299f53cf41f750eb56d8f8282c2fb75208f/t_ai_project-0.1.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "1c71ef9f36ee7e26adb00abaa6d2f576f8e61b5e9863f60604ebfc41b996073f",
"md5": "bc20825502e8f1f2aa452f7d56f581a0",
"sha256": "ddc6659a39a33b4446847d26656b61873d4f98edb4468a629469805b76fc8b00"
},
"downloads": -1,
"filename": "t_ai_project-0.1.0.tar.gz",
"has_sig": false,
"md5_digest": "bc20825502e8f1f2aa452f7d56f581a0",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 405219,
"upload_time": "2025-07-28T16:27:51",
"upload_time_iso_8601": "2025-07-28T16:27:51.862008Z",
"url": "https://files.pythonhosted.org/packages/1c/71/ef9f36ee7e26adb00abaa6d2f576f8e61b5e9863f60604ebfc41b996073f/t_ai_project-0.1.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-07-28 16:27:51",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "yourusername",
"github_project": "cortana-agent",
"github_not_found": true,
"lcname": "t-ai-project"
}
