Trollmaker

A fun and structured package that generates light-hearted, humorous roasts based on a public username or handle from social media. The output is formatted consistently with a roast, a playful insult, and a balanced compliment—all while ensuring a non-offensive, entertaining experience.

Installation

Install the package via pip:

pip install trollmaker

Usage

Basic Usage (Default LLM7)

from trollmaker import trollmaker

response = trollmaker(user_input="username_to_roast")
print(response)

Custom LLM Integration

By default, the package uses ChatLLM7 (from langchain_llm7). You can easily replace it with other LLMs like OpenAI, Anthropic, or Google Vertex AI.

Example: Using OpenAI

from langchain_openai import ChatOpenAI
from trollmaker import trollmaker

llm = ChatOpenAI()
response = trollmaker(user_input="username_to_roast", llm=llm)
print(response)

Example: Using Anthropic

from langchain_anthropic import ChatAnthropic
from trollmaker import trollmaker

llm = ChatAnthropic()
response = trollmaker(user_input="username_to_roast", llm=llm)
print(response)

Example: Using Google Vertex AI

from langchain_google_genai import ChatGoogleGenerativeAI
from trollmaker import trollmaker

llm = ChatGoogleGenerativeAI()
response = trollmaker(user_input="username_to_roast", llm=llm)
print(response)

Parameters

How It Works

1. Takes a username/handle as input.
2. Uses LLM7 (or a custom LLM) to generate a structured roast.
3. Ensures the output follows a consistent format (roast + insult + compliment).
4. Returns a humorous yet non-offensive response.

Rate Limits & API Key

• Default LLM7 Free Tier is sufficient for most use cases.
• For higher rate limits, set LLM7_API_KEY via environment variable or pass it directly:

  trollmaker(user_input="username", api_key="your_api_key")
  

• Get a free LLM7 API key at https://token.llm7.io/.

Contributing & Issues

For bugs, feature requests, or questions, open an issue here: 🔗 https://github.com/chigwell/trollmaker/issues

Author

👤 Eugene Evstafev 📧 hi@euegne.plus 🔗 GitHub: chigwell