pica-langchain

A Python package for integrating Pica with LangChain.

Full Documentation: https://docs.picaos.com/sdk/langchain

Installation

pip install pica-langchain

Usage

The PicaClientOptions class allows you to configure the Pica client with the following options:

The create_pica_agent function allows customizing the following parameters:

Quick Start

from langchain_openai import ChatOpenAI
from langchain.agents import AgentType
from pica_langchain import PicaClient, create_pica_agent
from pica_langchain.models import PicaClientOptions

# Initialize the Pica client
pica_client = PicaClient(
    secret="your-pica-secret",
    options=PicaClientOptions(
        # server_url="https://my-self-hosted-server.com",
        # identity_type="user"
        # identity="user-id",
        # authkit=True,
        # actions=[""], # Initialize specific action ids (e.g. ["conn_mod_def::F_JeJ_A_TKg::cc2kvVQQTiiIiLEDauy6zQ"])
        # permissions="read", # Filter actions by permission level
        
        connectors=["*"] # Initialize all available connections for this example
    )
)

pica_client.initialize()

# Create a LangChain agent with Pica tools
llm = ChatOpenAI(
    temperature=0, 
    model="gpt-4.1"
)

# Create an agent with Pica tools
agent = create_pica_agent(
    client=pica_client,
    llm=llm,
    agent_type=AgentType.OPENAI_FUNCTIONS,
    # return_intermediate_steps=True, # Optional: Return intermediate steps

    # Optional: Custom system prompt to append
    system_prompt="Always start your response with `Pica works like ✨\n`"
)

# Use the agent
result = agent.invoke({
    "input": (
            "Star the picahq/pica repo in github. "
            "Then, list 5 of the repositories that I have starred in github."
    )
})

print(result)

Using Individual Tools

from langchain.agents import AgentType, initialize_agent
from langchain_openai import ChatOpenAI
from pica_langchain import PicaClient, create_pica_tools

# Initialize the Pica client
pica_client = PicaClient(secret="your-pica-secret")

pica_client.initialize()

# Create Pica tools
tools = create_pica_tools(pica_client)

# Create a custom agent with the tools
llm = ChatOpenAI(temperature=0, model="gpt-4.1")
agent = initialize_agent(
    tools,
    llm,
    agent=AgentType.OPENAI_FUNCTIONS
)

# Use the agent
result = agent.run("What actions are available in Gmail?")
print(result)

Using Model Context Protocol (MCP) Tools

The SDK supports integration with Model Context Protocol (MCP) servers, allowing you to connect to external tool providers via the MCP protocol.

import asyncio
from langchain.agents import AgentType
from langchain_openai import ChatOpenAI
from pica_langchain import PicaClient, create_pica_agent
from pica_langchain.models import PicaClientOptions

# Configure MCP servers
mcp_options = {
    "math": {
        "command": "python",
        "args": ["./path/to/math_server.py"],
        "transport": "stdio",
    },
    "weather": {
        "url": "http://localhost:8000/sse",
        "transport": "sse",
    }
}

async def main():
    # Create client with async initialization
    pica_client = await PicaClient.create(
        secret="your-pica-secret",
        options=PicaClientOptions(
            connectors=["*"],  # Initialize all available connections
            mcp_options=mcp_options,  # Add MCP server options
        ),
    )

    pica_client.initialize()

    # Create an agent with both Pica and MCP tools
    llm = ChatOpenAI(temperature=0, model="gpt-4.1")
    agent = create_pica_agent(
        client=pica_client,
        llm=llm,
        agent_type=AgentType.OPENAI_FUNCTIONS,
    )

    # Use both Pica platform actions and MCP tools
    result = await agent.ainvoke({
        "input": "Calculate 25 * 17, then check the weather in New York, finally list all connectors Pica supported"
    })
    
    print(result["output"])

if __name__ == "__main__":
    asyncio.run(main())

Development

Setup

1. Clone the repository:

git clone https://github.com/yourusername/pica-langchain.git
cd pica-langchain

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate

3. Create a connection on Pica:

1. Create an account on app.picaos.com.
2. Navigate to the "My Connections" tab and create the required connection.
3. Retrieve your API Key from the "API Keys" section.

4. Export required environment variables:

export PICA_SECRET="your-pica-secret"
export OPENAI_API_KEY="your-openai-api-key"

5. Install development dependencies:

pip install -e ".[dev]"

Running Tests

pytest

Logging

The Pica LangChain SDK uses the logging module to log messages. The log level can be set using the PICA_LOG_LEVEL environment variable.

The following log levels are available:

• debug
• info
• warning
• error
• critical

export PICA_LOG_LEVEL="debug"

Examples

Examples can be found in the examples directory.

> python3 examples/use_with_langchain.py # LangChain agent example

License

This project is licensed under the GNU General Public License v3.0.