My Magic Pencil – LangChain Integration

LangChain tool for generating My Magic Pencil live visual lessons (drawing + voice) from text prompts.

Installation

pip install mymagicpencil-langchain

Environment Variables

Quick Start

import os
from mymagicpencil_langchain import GenerateVisualLessonTool

os.environ["MYMAGICPENCIL_API_KEY"] = "your-api-key"

tool = GenerateVisualLessonTool()
result = tool.run("Explain how photosynthesis works")
print(result)

Or pass the API key directly:

tool = GenerateVisualLessonTool(api_key="your-api-key")
result = tool.run("Explain Newton's laws of motion")
print(result)

Parameters

The GenerateVisualLessonTool accepts the following parameters when invoked:

Usage with LangChain Agents

import os
from langchain.agents import initialize_agent, AgentType
from langchain_openai import ChatOpenAI
from mymagicpencil_langchain import GenerateVisualLessonTool

os.environ["MYMAGICPENCIL_API_KEY"] = "your-mmp-api-key"
os.environ["OPENAI_API_KEY"] = "your-openai-api-key"

llm = ChatOpenAI(model="gpt-4", temperature=0)
tools = [GenerateVisualLessonTool()]

agent = initialize_agent(
    tools,
    llm,
    agent=AgentType.OPENAI_FUNCTIONS,
    verbose=True,
)

response = agent.run("Create a visual lesson explaining how black holes form")
print(response)

API Reference

GenerateVisualLessonTool

• Class: mymagicpencil_langchain.GenerateVisualLessonTool
• Inherits: langchain_core.tools.BaseTool
• Constructor Args:
  • api_key (Optional[str]): API key. Falls back to MYMAGICPENCIL_API_KEY env var.
• Methods:
  • _run(prompt, ...) → str: Calls the My Magic Pencil API and returns a result string with the embed URL and session ID.

GenerateVisualLessonInput

• Class: mymagicpencil_langchain.tool.GenerateVisualLessonInput
• Inherits: pydantic.BaseModel
• Purpose: Pydantic schema used for LangChain argument validation.

License

MIT