Using Composio With LiteLLM

Star A Repository on Github

In this example, we will use LiteLLM to star a repository on Github using Composio Tools

Install Packages

Python

$  pip install composio_openai litellm

Import Libraries, Initialize ComposioToolSet & Create Router

Python

$ from composio_openai import ComposioToolSet, App
> from litellm.router import Router
> 
> toolset = ComposioToolSet()
> 
> router = Router(
>     model_list=[
>         {
>             "model_name": "anthropic/claude-3-5",
>             "litellm_params": {
>                 "model": "claude-3-opus-20240229",
>                 "api_key": "<your-anthropic-api-key>",
>             },
>         }
>     ]
> )

Connect Your GitHub Account

You need to have an active GitHub Integration. Learn how to do this here

$ composio login 
> composio add github

Don’t forget to set your COMPOSIO_API_KEY and OPENAI_API_KEY in your environment variables.

Get All Github Tools

You can get all the tools for a given app as shown below, but you can get specific actions and filter actions using usecase & tags. Learn more here

Python

1 tools = toolset.get_tools(apps=[App.GITHUB])

Configure Router

Python

1 response = router.completion(
2     model="anthropic/claude-3-5",
3     messages=[
4         {"role": "user", "content": "Star the repository composiohq/composio on github"},
5     ],
6     tools=tools,
7     tool_choice="auto",
8 )

Execute Tool Calls

Python

1 result = toolset.handle_tool_calls(response=response)
2 print(result)

$	from composio_openai import ComposioToolSet, App
>	from litellm.router import Router
>
>	toolset = ComposioToolSet()
>
>	router = Router(
>	model_list=[
>	{
>	"model_name": "anthropic/claude-3-5",
>	"litellm_params": {
>	"model": "claude-3-opus-20240229",
>	"api_key": "<your-anthropic-api-key>",
>	},
>	}
>	]
>	)

1	response = router.completion(
2	model="anthropic/claude-3-5",
3	messages=[
4	{"role": "user", "content": "Star the repository composiohq/composio on github"},
5	],
6	tools=tools,
7	tool_choice="auto",
8	)

1	result = toolset.handle_tool_calls(response=response)
2	print(result)