zerogpt 1.3.0

Python client for interacting with the ZeroGPT API and generating images.

ZeroGPT

Documentation

ZeroGPT is a Python library for interacting with AI APIs, providing capabilities for text and image generation.

Features

• Text generation using various models
• Image creation based on textual descriptions
• Support for "uncensored" mode for more unrestricted responses
• Optimized memory and data handling
• Streamed data support
• Secure request authentication

Installation

pip install zerogpt

Usage

Client Initialization

from zerogpt import Client

client = Client()

Text Generation

# Simple request
response = client.send_message("Hi, how are you?")

# Request with instruction
response = client.send_message(
    "Tell me about space",
    instruction="You are an astronomy expert"
)

# Using "uncensored" mode
response = client.send_message(
    "Explain a complex topic",
    uncensored=True
)

# Using "think" mode (deeper reasoning)
response = client.send_message(
    "Solve a difficult math problem",
    think=True
)

# With context
messages=[
    {"role": "user", "content": "Hi"},
    {"role": "assistant", "content": "Hello!"}
]
response = client.send_message(
    messages,
    think=True
)

Image Generation

# Create image
result = client.create_image(
    prompt="anime neko girl",
    samples=1,
    resolution=(768, 512),
    seed=-1,
    steps=50
)

# Get generated image
image = client.get_image(result['data']['request_id'])

# Save image
image.download(['path/to/save/image.png'])

# View image
image.open()

Image to Prompt

from zerogpt.utils.tools import image_to_prompt

resp = image_to_prompt('path/to/image.png')

Working with Dummy Context[^1]

from zerogpt.utils.prompt import Dummy

# Create context
dummy = Dummy()
dummy.create(messages=[
    {"role": "user", "content": "Hi"},
    {"role": "assistant", "content": "Hello!"}
])

# Also possible for image generation
dummy = Dummy()
dummy.create(prompt='neko girl', steps=100)

# Save context
dummy.save("context.bin")

# Load context
dummy.load("context.bin")

# Use instead of messages:
# client.send_message(dummy)
# or
# client.create_image(dummy)

Parameters

send_message

• input (str or list): Text prompt or list of messages
• instruction (str, optional): System instruction
• think (bool, optional): Use model with deeper reasoning
• uncensored (bool, optional): Use unrestricted mode

create_image

• prompt (str): Description of the desired image
• samples (int, optional): Number of samples
• resolution (tuple, optional): Image resolution (width, height)
• seed (int, optional): Seed for reproducibility
• steps (int, optional): Number of generation steps
• negative_prompt (str, optional): Description of undesired elements

Security

The library uses HMAC-SHA256 to sign requests and ensure secure data transmission. All requests are authenticated using timestamps to prevent replay attacks.

Requirements

• Python 3.8+

License

MIT License Copyright (c) 2025 RedPiar

Author

RedPiar

[^1]: Dummy is used to compress context and data in general, very useful for systems with low RAM. It can also be saved for even greater memory efficiency!