OpenRouter dolphin API Quickstart Guide

Base URL and Authentication

The base URL for all requests is https://api.openrouterdolphin.com/v1. You authenticate by passing your API key in the Authorization

Copycurl https://api.openrouterdolphin.com/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "List the top 10 libraries for Python web scraping."}]
  }'

Authentication is strict. We do not support multiple keys per account. Generating a new key immediately invalidates the previous one. This ensures clean usage tracking for your prepaid credits.

Endpoint: POST /v1/chat/completions

Send text data to generate responses using the uncensored model. This endpoint handles all standard chat completions. You define the model, messages, and parameters in the request body.

Copyfrom openai import OpenAI

client = OpenAI(base_url="https://api.openrouterdolphin.com/v1", api_key="YOUR_KEY")

resp = client.chat.completions.create(
    model="uncensored",
    messages=[{"role": "user", "content": "Summarise this thread without softening it."}],
)
print(resp.choices[0].message.content)

The API returns a standard JSON object containing the generated text. It does not support embeddings, images, or audio. Focus on text generation workflows. The model responds to prompts without filtering lawful adult or controversial topics.

Endpoint: GET /v1/models

Retrieve available models via a simple GET request. This endpoint confirms the model ID you must use.

Copyimport OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.openrouterdolphin.com/v1", apiKey: process.env.API_KEY });

const resp = await client.chat.completions.create({
  model: "uncensored",
  messages: [{ role: "user", content: "Draft a villain monologue for my game." }],
});
console.log(resp.choices[0].message.content);

The response lists the uncensored model. Use this exact ID in your completion requests. The API serves only this single model. There is no routing or model selection logic. You interact directly with our dedicated uncensored LLM instance.

Streaming and Usage Data

Enable streaming by setting stream: true. The API returns chunks via Server-Sent Events (SSE). Each chunk contains a portion of the response text.

Copystream = client.chat.completions.create(
    model="uncensored",
    messages=[{"role": "user", "content": "Tell the story in second person."}],
    stream=True,
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

The final chunk includes the full token usage statistics. You pay strictly for the output tokens generated. Errors and refusals do not incur charges. This structure allows real-time rendering of responses in your applications.

Parameters: Temperature, Top P, and Tools

Control generation behavior with standard parameters. Set temperature for randomness and top_p for nucleus sampling. Enable function calling by providing a tools array.

Use response_format with {"type": "json_object"} for strict JSON output. You can also set seed for reproducibility. The API supports presence_penalty and frequency_penalty for fine-tuning word usage. These parameters work exactly as they do in standard OpenAI clients.

Rate Limits and Request Size

Each API key allows 300 requests per minute. You can have up to 8 concurrent requests at any time. The maximum request body size is 8 MB. The context window supports 100,000 tokens for input and output combined.

Max output per request is 16,000 tokens. If you do not set max_tokens, the default is 2,048 tokens. Exceeding these limits returns a 429 error. Plan your batching strategy accordingly. Credits never expire, so you can accumulate usage over time without pressure.