SDKs & Libraries

Official GPTfake SDKs for accessing AI censorship monitoring data programmatically. Both wrap the GPTfake API and return the same metrics you see on each model page — for example the live ChatGPT refusal data — and across the Monitoring dashboard.

Official SDKs

Python SDK

Installation


pip install gptfake

Quick start


from gptfake import GPTfakeClient
 
client = GPTfakeClient(api_key="your-api-key")
 
# Get current metrics for a model
metrics = client.monitoring.get_metrics("chatgpt")
print(f"Censorship Rate: {metrics.censorship_rate}%")
print(f"Bias Score: {metrics.bias_score}")
print(f"Transparency: {metrics.transparency_score}/100")
 
# Compare multiple models
comparison = client.monitoring.compare_models(
    models=["chatgpt", "claude", "gemini"]
)
for model in comparison:
    print(f"{model.name}: {model.censorship_rate}% censorship")
 
# Get historical data
history = client.monitoring.get_history("chatgpt", days=30)
for day in history:
    print(f"{day.date}: {day.censorship_rate}%")

Features

Async/await support with asyncio
Pandas DataFrame integration
Type hints included
Comprehensive documentation
Jupyter notebook friendly

JavaScript SDK

Installation


npm install gptfake

Quick start


import { GPTfakeClient } from 'gptfake';
 
const client = new GPTfakeClient({
  apiKey: 'your-api-key'
});
 
// Get current metrics
const metrics = await client.monitoring.getMetrics('chatgpt');
console.log(`Censorship Rate: ${metrics.censorshipRate}%`);
 
// Compare models
const comparison = await client.monitoring.compareModels([
  'chatgpt', 'claude', 'gemini'
]);
 
// Get historical data
const history = await client.monitoring.getHistory('chatgpt', { days: 30 });

Features

Full TypeScript support
Promise-based API
Browser and Node.js compatible
Automatic retry logic
Comprehensive error handling

SDK comparison

Feature	Python	JavaScript
Installation	pip	npm
Type support	Type hints	TypeScript
Async	asyncio	Promise
Pandas integration	Yes	No
Browser support	No	Yes
Documentation	Comprehensive	Comprehensive

Common operations

Get model metrics


# Python
metrics = client.monitoring.get_metrics("chatgpt")


// JavaScript
const metrics = await client.monitoring.getMetrics('chatgpt');

Compare models


# Python
comparison = client.monitoring.compare_models([
    "chatgpt", "claude", "gemini", "mistral", "qwen"
])


// JavaScript
const comparison = await client.monitoring.compareModels([
  'chatgpt', 'claude', 'gemini', 'mistral', 'qwen'
]);

Get historical data


# Python
history = client.monitoring.get_history("chatgpt", days=30)


// JavaScript
const history = await client.monitoring.getHistory('chatgpt', { days: 30 });

Export data


# Python
data = client.research.export_data(
    models=["chatgpt", "claude"],
    format="csv",
    days=90
)


// JavaScript
const data = await client.research.exportData({
  models: ['chatgpt', 'claude'],
  format: 'csv',
  days: 90
});

Error handling


# Python
from gptfake.exceptions import (
    AuthenticationError,
    RateLimitError,
    ModelNotFoundError
)
 
try:
    metrics = client.monitoring.get_metrics("invalid-model")
except AuthenticationError:
    print("Check your API key")
except RateLimitError:
    print("Too many requests, please wait")
except ModelNotFoundError:
    print("Model not found")


// JavaScript
try {
  const metrics = await client.monitoring.getMetrics('invalid-model');
} catch (error) {
  if (error.code === 'AUTHENTICATION_ERROR') {
    console.log('Check your API key');
  } else if (error.code === 'RATE_LIMIT_ERROR') {
    console.log('Too many requests');
  }
}

CLI tool

In addition to the SDKs, we offer a command-line interface:


pip install gptfake-cli
 
# Get metrics
gptfake metrics chatgpt
 
# Compare models
gptfake compare chatgpt claude gemini
 
# Export data
gptfake export --format csv --days 30 --output data.csv

Contributing

Our SDKs are open source:

Python SDK: github.com/gptfake/gptfake-python
JavaScript SDK: github.com/gptfake/gptfake-js

We welcome contributions, bug reports, and feature requests.

Support

Tutorials — step-by-step guides
API reference — API documentation
Support — troubleshooting & contact
Email — [email protected]