Uncensored AI Integration Guide

Why Switch to an Uncensored AI Model?

Uncensored AI models remove the polite refusal layer that standard models apply to lawful but edgy topics. Standard LLMs often reject prompts about politics, adult themes, or creative fiction simply because they are trained to be safe. This behavior frustrates developers who need raw reasoning capabilities without artificial constraints.

Our API serves a model tuned to answer without these content refusals. You get the full output the model generates, including nuanced arguments or explicit creative content, provided it is lawful. This is not just about NSFW content; it is about removing bias and over-censorship from general-purpose tasks.

• True freedom for creative writing and roleplay scenarios.
• Better performance on security research and prompt injection testing.
• No hidden filters blocking specific political or social viewpoints.

Understanding the OpenAI-Compatible Standard

The OpenAI-compatible standard means your code talks to any provider using the same JSON structure. You send a message payload, and receive a standardized response object. This portability is the key advantage of modern LLM APIs.

Most developers already have code written for the OpenAI SDK. By using an OpenAI-compatible endpoint, you can swap the provider without touching your application logic. The API structure remains identical, ensuring a smooth transition.

Key Differences from Standard Providers

• Single model focus: No routing logic to slow down responses.
• Transparent pricing: Pay only for tokens used, no hidden tier fees.
• Direct access: Connect directly to the uncensored model without intermediaries.

Step 1: Install the OpenAI SDK

Install the official OpenAI SDK for your preferred programming language. This library handles the HTTP requests and JSON parsing for you. It is the industry standard for interacting with LLM APIs.

For Python, use pip. For Node.js, use npm. The library is lightweight and integrates easily into existing projects. Ensure you are using the latest version to support all features like streaming and tool calling.

Step 2: Configure the Base URL

Point your SDK client to our API endpoint. This is the only configuration change required to switch providers. The base URL directs your requests to our servers instead of OpenAI’s.

Use https://api.dolphinuncensored.com/v1 as your base URL. This ensures all requests go to the correct uncensored model. You can set this in your environment variables or directly in your client configuration.

Step 3: Initialize the Client

Create a client instance using your API key. The key authenticates your requests and tracks your usage. Pass the base URL to ensure the client knows where to send data.

The client handles retries and error codes automatically. This reduces boilerplate code in your application. You can initialize the client once and reuse it for multiple requests.

Step 4: Send Your First Completion

Send a chat completion request to test the integration. The model will process your prompt and return a response. This confirms that your configuration is correct.

Use the model ID uncensored to access the specific uncensored model. The response will include the generated text and token usage statistics. This is a standard request structure compatible with all OpenAI-compatible clients.

Step 5: Implement Streaming for Real-Time Output

Enable streaming to receive tokens as they are generated. This improves user experience by showing progress immediately. Long responses can take seconds to complete; streaming prevents the UI from freezing.

Set stream=True in your request. The SDK will yield chunks of data as they arrive. Each chunk contains part of the response text. This is essential for chat interfaces and real-time applications.

Handling Tokens and Context Windows

Our API supports a context window of 100,000 tokens. This allows for extensive conversations and large document processing. The maximum output per request is 16,000 tokens.

Monitor your token usage to manage costs efficiently. Input tokens are charged at $0.15 per million, and output tokens at $0.50 per million. Errors and refusals do not incur charges. This transparent pricing helps you predict expenses accurately.

Token Management Tips

• Keep conversation history concise to save tokens.
• Use system prompts to set context without repeating data.
• Monitor usage via the /v1/models endpoint.

Next Steps for Production

Integrate error handling and rate limiting into your application. Our API supports 300 requests per minute per key. Plan your traffic accordingly to avoid throttling.

Consider using function calling for structured data extraction. This feature allows the model to return JSON objects instead of plain text. It is useful for parsing data or triggering actions in your application.