What is dolphin AI and why use it
Dolphin AI is an uncensored large language model optimized for creative freedom and unrestricted dialogue. It strips away the conservative safety filters often found in mainstream commercial models. This makes it ideal for NSFW roleplay, unfiltered creative writing, and robust coding tasks where strict adherence to safety guidelines might interrupt the flow. The model is abliterated, meaning it has been specifically tuned to answer questions without refusal for lawful adult content.
Developers choose this path when standard APIs reject valid creative outputs or censor nuanced discussions. The Dolphin AI ecosystem provides a direct route to these capabilities. You get a raw, unvarnished text generation engine that prioritizes completion over politeness. This approach suits applications requiring high fidelity in character voice or unrestricted data processing.
Understanding the uncensored nature of the model
The uncensored nature of the model means it does not refuse lawful adult, fictional, or controversial topics. It operates without the typical guardrails that limit output diversity. This includes sexual content involving adults, which remains permissible as long as it does not involve minors. The model handles dark themes, profanity, and niche creative styles with equal ease.
Unlike generic aggregators that might route your request through multiple filtered layers, this implementation serves the base model directly. You receive the raw output without intermediate moderation steps. This transparency ensures that your application controls the final presentation. If you need raw data for analysis or creative exploration, this model delivers exactly what is asked without unexpected censorship.
Setting up your development environment
Start by installing the official OpenAI SDK for your preferred language. The setup requires minimal configuration changes. You need to define the base URL pointing to our endpoint. Then, provide your API key from the dashboard. This key authenticates your requests immediately.
- Install the SDK via pip, npm, or your package manager.
- Create a new client instance.
- Set the
base_urltohttps://api.dolphinllmmodel.com/v1. - Assign your API key to the
api_keyparameter.
This configuration aligns your environment with the OpenAI standard. You can now proceed to make requests using familiar method calls. The transition from other providers is seamless because the protocol remains consistent.
Connecting your openai compatible client
Connecting your client involves pointing your existing code to the new endpoint. The model ID is simply uncensored. This ID maps to the single model hosted on our servers. It is not GPT, Claude, or any other vendor's model. It is an open-weight model tuned for unrestricted output.
You can use any library that supports the OpenAI chat completions format. The request structure remains identical. You send a message array and receive a text response. This compatibility allows you to swap providers without rewriting your application logic. It is a robust solution for developers who want to avoid vendor lock-in while maintaining code simplicity.
Making your first uncensored request
Your first request demonstrates the model's responsiveness. You send a prompt through the chat/completions endpoint. The model processes the input and returns text. There are no hidden fees for this initial interaction. You only pay for the tokens consumed.
The pricing is transparent. Input tokens cost $0.15 per million, and output tokens cost $0.50 per million. Errors are free. Refusals are free. This structure ensures you only pay for value delivered. You can start with the trial credit if you want to test before committing. The trial gives you $0.50 valid for seven days.
Implementing streaming responses
Streaming responses provide a better user experience for long outputs. You enable streaming by setting the stream parameter to true. The API returns chunks of text as they are generated. This reduces latency perception for the end user.
The final chunk includes token usage statistics. This allows you to track costs in real-time. You can handle errors or interruptions gracefully. Streaming is essential for chat interfaces where instant feedback matters. It keeps the connection active and responsive. The model supports this feature natively.
Handling token limits and context windows
The context window supports up to 100,000 tokens combined for prompt and completion. This allows for extensive conversations or large document processing. You can set the maximum output to 16,000 tokens. If you do not specify max_tokens, the default is 2,048 tokens. This limit prevents runaway responses.
Monitoring token usage is crucial for cost control. You can adjust the max_tokens parameter to fit your needs. Requests exceeding the context window will fail. The request body is limited to 8 MB. This ensures efficient data transfer. You can maintain long-term memory by managing the conversation history manually.
Advanced features: tools and JSON mode
Advanced features include function calling and JSON mode. You can define tools in your request. The model responds with structured data that your code can execute. This enables complex workflows and agent-like behavior. You can also force JSON output using response_format. This ensures predictable parsing for your application.
Parameters like temperature and top_p allow fine-tuning of creativity. Use seed for reproducible results. presence_penalty and frequency_penalty help control repetition. These tools give you precise control over the model's behavior. You can optimize for accuracy or creativity as needed.
Troubleshooting common integration issues
Common issues often stem from incorrect base URLs or missing API keys. Verify that your endpoint matches https://api.dolphinllmmodel.com/v1. Check that your key is active and has sufficient credit. Rate limits are 300 requests per minute and 8 concurrent requests per key. Exceeding these limits returns a 429 error.
If you encounter errors, check the request body size. It must not exceed 8 MB. Ensure that your JSON is valid. Streaming issues often arise from improper chunk handling. Review the final chunk for usage data. Support is available via the Support page for billing issues like double charges. Credit never expires, so your balance is safe.
