Unfiltered AI Chat: A Step-by-Step Integration Guide

Why Unfiltered AI Chat Matters for Developers

Developers building roleplay apps often face silent refusals that break the immersion of their user experience. Standard models often downgrade content quality or block requests based on vague 'vibe' checks rather than hard policy violations. An unfiltered AI chat solution removes these arbitrary gatekeepers, allowing your application to handle diverse, mature, or controversial topics without constant friction. This reliability is critical for content pipelines where consistency matters more than generic safety.

Our infrastructure is purpose-built to serve requests for adult and unrestricted content. We do not down-rank your traffic for being 'NSFW'. Instead, we provide a direct path to an uncensored model that answers questions and generates stories based on context, not on a hidden moderation layer. This approach ensures your users get the raw, unpolished creativity they expect from dedicated chatbots.

Setting Up Your Environment

Start by installing the official OpenAI SDK for your preferred language, as our API follows the standard chat-completions protocol. You will need to configure your environment variables to store your API key securely. The setup process is straightforward because we rely on existing industry standards.

Install the library using your package manager. For Python, use pip install openai. For Node.js, use npm install openai. Once installed, import the client and initialize it with your base URL. This step ensures that all subsequent calls route correctly to our servers. Keep your key private and never expose it in client-side code unless you are using a proxy.

Configuring the Base URL

Point your client to HTTPS://api.storyunfiltered.com/v1 to access the uncensored endpoint. This base URL replaces the default OpenAI endpoint in your SDK configuration. By changing this single variable, you switch your application from the standard GPT models to our dedicated uncensored large language model. The model ID you must specify in your requests is "uncensored". This identifier tells the server which specific instance to route your traffic to.

Verify the configuration by sending a simple test request. Check the GET /v1/models endpoint to confirm the model is available. The response should list "uncensored" as an available model. This verification step ensures your environment is correctly wired before you begin building complex features. It also confirms that the API key you generated is active and linked to your prepaid credit.

Authenticating with Your API Key

Pass your API key in the Authorization header as a Bearer token. We support key generation via Google login or email and password. The key is displayed immediately after creation, so copy it to your environment variables. Each account holds one active key; generating a new one replaces the previous key instantly. This simplifies key rotation and reduces the risk of stale credentials lingering in your systems.

No phone number is required for signup, and privacy is maintained by storing only your email. Prompts are not used for training your data, which is crucial for enterprise or privacy-conscious applications. The key does not expire, but the prepaid credit attached to it does not have a hard expiration date either. Errors and refusals do not consume your balance, so you can test freely without worrying about unexpected charges.

Sending Your First Request

Construct a standard POST request to /v1/chat/completions with the model set to "uncensored". Include your system prompt to define the persona, and add user messages to drive the conversation. The context window supports 100,000 tokens for both prompt and completion combined. This allows you to pass large documents or long conversation histories without immediate truncation. Ensure your payload stays under the 8 MB request body limit.

Set the temperature and top_p parameters to control creativity. A higher temperature yields more diverse outputs, which is ideal for roleplay. A lower temperature ensures consistency for factual or structured tasks. The model supports JSON mode via response_format, allowing you to enforce structured outputs for data extraction tasks. This flexibility makes the API suitable for both casual chat and production-grade integrations.

Handling Streaming Responses

Enable streaming by setting stream to true in your request. The API returns Server-Sent Events (SSE) containing partial responses as they are generated. This improves the perceived latency for end-users, making the chat feel responsive even for long outputs. Each chunk contains delta content that your application must concatenate to form the final response.

The final chunk in the stream includes token usage data. Use this data to track your credit consumption accurately. We charge $0.15 per 1M input tokens and $0.50 per 1M output tokens. Billing is based on real usage, so partial responses are charged proportionally. This transparency allows you to implement cost controls directly in your application logic.

Implementing Tool Calling

Our API supports function calling, allowing the model to invoke external tools based on user intent. Define your functions in the tools parameter with a name, description, and parameters schema. The model will return a response with tool_calls if it determines a function is needed. You must then execute the function and send the result back in a subsequent message.

This feature enables complex workflows like weather lookups, database queries, or web scraping. The uncensored model handles tool definitions with the same precision as standard models. However, note that we do not offer embeddings or fine-tuning. If your application relies on vector search or custom model training, this API may not be the right fit. Stick to direct chat-completions for the best results.

Optimizing for Cost and Speed

Maximize the utility of your prepaid credit by minimizing unnecessary API calls. Use the seed parameter for deterministic outputs during testing. Adjust the max_tokens limit to prevent the model from generating excessively long responses that drain your balance. The default output limit is 2,048 tokens, but you can increase it up to 16,000 tokens per request. This is useful for generating long stories or detailed reports in a single call.

Monitor your request rate to stay within the 300 requests per minute limit. If you exceed this, you will receive a 429 error. Implement exponential backoff in your client code to handle retries gracefully. Since errors are free, you can afford to retry failed requests without penalty. This safety net ensures reliability even during peak usage periods.

Troubleshooting Common Errors

If you receive a 401 error, verify that your API key is correct and active. Check that the base URL matches HTTPS://api.storyunfiltered.com/v1 exactly. A 404 error usually indicates a typo in the endpoint path. Ensure you are using the correct HTTP method, which is POST for chat completions. If the error persists, regenerate your key from the dashboard.

A 429 error means you have exceeded the rate limit of 300 requests per minute. Wait for the window to reset or reduce your request frequency. If you encounter a 413 error, your request body exceeds the 8 MB limit. Simplify your payload or reduce the context window size. These errors are standard for REST APIs and can be resolved with careful debugging.