# Pydantic AI LiteLLM
A LiteLLM model integration for the [Pydantic AI](https://ai.pydantic.dev/) framework, enabling access to 100+ LLM providers through a unified interface.
## Features
- **Universal LLM Access**: Connect to 100+ LLM providers (OpenAI, Anthropic, Cohere, Bedrock, Azure, and many more) via LiteLLM
- **Full Pydantic AI Integration**: Complete support for tool calling, streaming, structured outputs, and all Pydantic AI features
- **Type Safety**: Fully typed with comprehensive type hints
- **Async/Await Support**: Built for modern async Python applications
- **Flexible Configuration**: Support for custom API endpoints, headers, and provider-specific settings
## Installation
```bash
pip install pydantic-ai-litellm
```
## Quick Start
```python
import asyncio
from pydantic_ai import Agent
from pydantic_ai_litellm import LiteLLMModel
# Initialize with any LiteLLM-supported model
model = LiteLLMModel(
model_name="gpt-4", # or claude-3-opus, gemini-pro, etc.
api_key="your-api-key" # will also check environment variables
)
# Create an agent
agent = Agent(model=model)
# Run inference
async def main():
result = await agent.run("What is the capital of France?")
print(result.output)
asyncio.run(main())
```
## Supported Providers
This library supports all providers available through LiteLLM, including:
- **OpenAI**: GPT-4, GPT-3.5, o1, etc.
- **Anthropic**: Claude 3 (Opus, Sonnet, Haiku)
- **Google**: Gemini Pro, Gemini Flash
- **AWS Bedrock**: Claude, Titan, Cohere models
- **Azure OpenAI**: All Azure-hosted models
- **Cohere**: Command, Command R+
- **Mistral AI**: Mistral 7B, 8x7B, Large
- **And 90+ more providers**
See the [LiteLLM providers documentation](https://docs.litellm.ai/docs/providers) for the complete list.
## Advanced Usage
### Custom API Endpoints
```python
model = LiteLLMModel(
model_name="custom-model",
api_base="https://your-custom-endpoint.com/v1",
api_key="your-api-key",
custom_llm_provider="openai" # specify provider format
)
```
### Tool Calling
```python
from pydantic_ai import Agent
from pydantic_ai_litellm import LiteLLMModel
def get_weather(location: str) -> str:
"""Get weather for a location."""
return f"It's sunny in {location}"
model = LiteLLMModel("gpt-4")
agent = Agent(model=model, tools=[get_weather])
result = await agent.run("What's the weather in Paris?")
```
### Streaming
```python
async with agent.run_stream("Write a poem about AI") as stream:
async for text in stream.stream_text(delta=True):
print(text, end="", flush=True)
```
### Structured Output
```python
from pydantic import BaseModel
class Person(BaseModel):
name: str
age: int
occupation: str
agent = Agent(model=model, output_type=Person)
result = await agent.run("Generate a person profile")
print(result.output.name) # Typed as Person
```
## Configuration
You can configure the model with various settings:
```python
from pydantic_ai_litellm import LiteLLMModelSettings
settings: LiteLLMModelSettings = {
'temperature': 0.7,
'max_tokens': 1000,
'litellm_api_key': 'your-key',
'litellm_api_base': 'https://custom-endpoint.com',
'extra_headers': {'Custom-Header': 'value'}
}
model = LiteLLMModel("gpt-4", settings=settings)
```
## Requirements
- Python 3.13+
- `pydantic-ai-slim>=0.6.2`
- `litellm>=1.75.5`
## Contributing
Contributions are welcome! Please feel free to submit issues and pull requests.
## License
MIT License - see LICENSE file for details.
## Examples
See the `examples/` directory for complete working examples:
- **Quick Start** (`examples/01_quick_start.py`) - Basic usage
- **Custom Endpoints** (`examples/02_custom_endpoints.py`) - Using custom API endpoints
- **Tool Calling** (`examples/03_tool_calling.py`) - Functions as AI tools
- **Streaming** (`examples/04_streaming.py`) - Real-time text streaming
- **Structured Output** (`examples/05_structured_output.py`) - Typed responses with Pydantic
- **Configuration** (`examples/06_configuration.py`) - Model settings and parameters
Each example includes error handling and can be run independently with the appropriate API keys.
## Links
- [Pydantic AI Documentation](https://ai.pydantic.dev/)
- [LiteLLM Documentation](https://docs.litellm.ai/)
- [GitHub Repository](https://github.com/mochow13/pydantic-ai-litellm)
Raw data
{
"_id": null,
"home_page": null,
"name": "pydantic-ai-litellm",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.13",
"maintainer_email": "Mottakin <md.mottakin.chowdhury@gmail.com>",
"keywords": "pydantic, ai, llm, litellm, openai, claude, gemini, machine-learning",
"author": null,
"author_email": "Mottakin <md.mottakin.chowdhury@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/26/ca/0dee552d69a215d9d6b93e15a6e189e7b101a8905ccd11f4c42ce13b41ed/pydantic_ai_litellm-0.2.0.tar.gz",
"platform": null,
"description": "# Pydantic AI LiteLLM\n\nA LiteLLM model integration for the [Pydantic AI](https://ai.pydantic.dev/) framework, enabling access to 100+ LLM providers through a unified interface.\n\n## Features\n\n- **Universal LLM Access**: Connect to 100+ LLM providers (OpenAI, Anthropic, Cohere, Bedrock, Azure, and many more) via LiteLLM\n- **Full Pydantic AI Integration**: Complete support for tool calling, streaming, structured outputs, and all Pydantic AI features\n- **Type Safety**: Fully typed with comprehensive type hints\n- **Async/Await Support**: Built for modern async Python applications\n- **Flexible Configuration**: Support for custom API endpoints, headers, and provider-specific settings\n\n## Installation\n\n```bash\npip install pydantic-ai-litellm\n```\n\n## Quick Start\n\n```python\nimport asyncio\nfrom pydantic_ai import Agent\nfrom pydantic_ai_litellm import LiteLLMModel\n\n# Initialize with any LiteLLM-supported model\nmodel = LiteLLMModel(\n model_name=\"gpt-4\", # or claude-3-opus, gemini-pro, etc.\n api_key=\"your-api-key\" # will also check environment variables\n)\n\n# Create an agent\nagent = Agent(model=model)\n\n# Run inference\nasync def main():\n result = await agent.run(\"What is the capital of France?\")\n print(result.output)\n\nasyncio.run(main())\n```\n\n## Supported Providers\n\nThis library supports all providers available through LiteLLM, including:\n\n- **OpenAI**: GPT-4, GPT-3.5, o1, etc.\n- **Anthropic**: Claude 3 (Opus, Sonnet, Haiku)\n- **Google**: Gemini Pro, Gemini Flash\n- **AWS Bedrock**: Claude, Titan, Cohere models\n- **Azure OpenAI**: All Azure-hosted models\n- **Cohere**: Command, Command R+\n- **Mistral AI**: Mistral 7B, 8x7B, Large\n- **And 90+ more providers**\n\nSee the [LiteLLM providers documentation](https://docs.litellm.ai/docs/providers) for the complete list.\n\n## Advanced Usage\n\n### Custom API Endpoints\n\n```python\nmodel = LiteLLMModel(\n model_name=\"custom-model\",\n api_base=\"https://your-custom-endpoint.com/v1\",\n api_key=\"your-api-key\",\n custom_llm_provider=\"openai\" # specify provider format\n)\n```\n\n### Tool Calling\n\n```python\nfrom pydantic_ai import Agent\nfrom pydantic_ai_litellm import LiteLLMModel\n\ndef get_weather(location: str) -> str:\n \"\"\"Get weather for a location.\"\"\"\n return f\"It's sunny in {location}\"\n\nmodel = LiteLLMModel(\"gpt-4\")\nagent = Agent(model=model, tools=[get_weather])\n\nresult = await agent.run(\"What's the weather in Paris?\")\n```\n\n### Streaming\n\n```python\nasync with agent.run_stream(\"Write a poem about AI\") as stream:\n async for text in stream.stream_text(delta=True):\n print(text, end=\"\", flush=True)\n```\n\n### Structured Output\n\n```python\nfrom pydantic import BaseModel\n\nclass Person(BaseModel):\n name: str\n age: int\n occupation: str\n\nagent = Agent(model=model, output_type=Person)\nresult = await agent.run(\"Generate a person profile\")\nprint(result.output.name) # Typed as Person\n```\n\n## Configuration\n\nYou can configure the model with various settings:\n\n```python\nfrom pydantic_ai_litellm import LiteLLMModelSettings\n\nsettings: LiteLLMModelSettings = {\n 'temperature': 0.7,\n 'max_tokens': 1000,\n 'litellm_api_key': 'your-key',\n 'litellm_api_base': 'https://custom-endpoint.com',\n 'extra_headers': {'Custom-Header': 'value'}\n}\n\nmodel = LiteLLMModel(\"gpt-4\", settings=settings)\n```\n\n## Requirements\n\n- Python 3.13+\n- `pydantic-ai-slim>=0.6.2`\n- `litellm>=1.75.5`\n\n## Contributing\n\nContributions are welcome! Please feel free to submit issues and pull requests.\n\n## License\n\nMIT License - see LICENSE file for details.\n\n## Examples\n\nSee the `examples/` directory for complete working examples:\n\n- **Quick Start** (`examples/01_quick_start.py`) - Basic usage\n- **Custom Endpoints** (`examples/02_custom_endpoints.py`) - Using custom API endpoints \n- **Tool Calling** (`examples/03_tool_calling.py`) - Functions as AI tools\n- **Streaming** (`examples/04_streaming.py`) - Real-time text streaming\n- **Structured Output** (`examples/05_structured_output.py`) - Typed responses with Pydantic\n- **Configuration** (`examples/06_configuration.py`) - Model settings and parameters\n\nEach example includes error handling and can be run independently with the appropriate API keys.\n\n## Links\n\n- [Pydantic AI Documentation](https://ai.pydantic.dev/)\n- [LiteLLM Documentation](https://docs.litellm.ai/)\n- [GitHub Repository](https://github.com/mochow13/pydantic-ai-litellm)\n",
"bugtrack_url": null,
"license": null,
"summary": "LiteLLM model integration for Pydantic AI framework - access 100+ LLM providers through a unified interface",
"version": "0.2.0",
"project_urls": {
"Bug Tracker": "https://github.com/mochow13/pydantic-ai-litellm/issues",
"Documentation": "https://github.com/mochow13/pydantic-ai-litellm#readme",
"Homepage": "https://github.com/mochow13/pydantic-ai-litellm",
"Repository": "https://github.com/mochow13/pydantic-ai-litellm"
},
"split_keywords": [
"pydantic",
" ai",
" llm",
" litellm",
" openai",
" claude",
" gemini",
" machine-learning"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "24290607c4d68b05624165f747257c305e6f56fa7f4bf41b1fd6248ebd947bb3",
"md5": "8a97c27b7f4dab80aa0f68a5d1db47a7",
"sha256": "6a7a7b5a691cd6fd36092d7c6b2d8aff7f5192c3117c86922ebbd6b35ebc8410"
},
"downloads": -1,
"filename": "pydantic_ai_litellm-0.2.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "8a97c27b7f4dab80aa0f68a5d1db47a7",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.13",
"size": 9060,
"upload_time": "2025-08-30T23:26:38",
"upload_time_iso_8601": "2025-08-30T23:26:38.066635Z",
"url": "https://files.pythonhosted.org/packages/24/29/0607c4d68b05624165f747257c305e6f56fa7f4bf41b1fd6248ebd947bb3/pydantic_ai_litellm-0.2.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "26ca0dee552d69a215d9d6b93e15a6e189e7b101a8905ccd11f4c42ce13b41ed",
"md5": "5ff363ce2e8067114a3605789dacc47b",
"sha256": "caa54b08018618088d205e9c9a03616ac0898a1cc0b8363cbff94ba739bec30e"
},
"downloads": -1,
"filename": "pydantic_ai_litellm-0.2.0.tar.gz",
"has_sig": false,
"md5_digest": "5ff363ce2e8067114a3605789dacc47b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.13",
"size": 8899,
"upload_time": "2025-08-30T23:26:39",
"upload_time_iso_8601": "2025-08-30T23:26:39.522574Z",
"url": "https://files.pythonhosted.org/packages/26/ca/0dee552d69a215d9d6b93e15a6e189e7b101a8905ccd11f4c42ce13b41ed/pydantic_ai_litellm-0.2.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-30 23:26:39",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "mochow13",
"github_project": "pydantic-ai-litellm",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "pydantic-ai-litellm"
}
JSON
