Dolphin series ai API Documentation

Authentication and base URL

Send your API key in the Authorization header to access the inference endpoint. We do not route between vendors; you connect directly to our single model server. The base URL is fixed for all requests, and you must include the model identifier "uncensored" in every payload. Your key is generated immediately after signing up with Google or an email address. There is no phone number verification required to start.

Keep your key secure, as it tracks your prepaid credit usage. Errors and refusals do not consume tokens, so failed requests are free. You can generate a new key at any time, which immediately invalidates the previous one.

Chat completions endpoint

Post your conversation history to the chat completions endpoint to receive text responses. This endpoint handles standard message arrays and returns generated tokens based on your parameters. It does not support embeddings, audio, or image generation. You must specify the model ID as uncensored in the request body.

• Supports streaming via Server-Sent Events.
• Accepts standard parameters like temperature and top_p.
• Returns usage statistics in the final chunk when streaming.

Requests are limited to an 8 MB body size. If you exceed the rate limit, you will receive a 429 error. Ensure your client handles retries appropriately without double-charging your credit.

Streaming responses

Enable streaming by setting stream: true in your request to receive tokens incrementally. This reduces perceived latency for long outputs and allows real-time display of generated text. The final chunk contains the complete usage statistics, including input and output token counts. You can parse these statistics to track your credit consumption accurately.

Streaming is ideal for chat interfaces where users expect immediate feedback. Note that the model does not guarantee a specific response time, as performance depends on server load. You must handle partial responses and potential network interruptions in your client code.

Function calling and tools

Define tools in your request to enable the model to call external functions. Use the tools parameter to specify function schemas, and set tool_choice to control whether the model decides or is forced to call a tool. The model returns structured JSON objects that you can parse and execute in your application. This works seamlessly with standard OpenAI-compatible SDKs.

Function calling is useful for extracting structured data or triggering actions based on user input. The model supports multiple tools in a single request. Ensure your function schemas are well-defined to minimize parsing errors in your client code.

Parameters: temperature, top_p, and seed

Adjust temperature to control randomness, with higher values producing more creative outputs. Use top_p to limit vocabulary diversity, and set seed for reproducible results. These parameters allow fine-tuning of the model's behavior for specific use cases. The default values are suitable for general-purpose text generation.

Setting a fixed seed ensures consistent outputs for testing and debugging. You can combine these parameters to achieve the desired balance between creativity and determinism. Experiment with different values to find the optimal settings for your application.

JSON mode and response formatting

Force the model to output valid JSON by setting response_format to {"type": "json_object"}. This ensures the output can be parsed directly by your application without additional validation. JSON mode is ideal for structured data extraction or API responses. The model will attempt to conform to the requested format, but complex schemas may still require client-side validation.

Use JSON mode in conjunction with function calling for robust tool integration. Ensure your client handles potential parsing errors gracefully. This feature reduces the need for complex regex patterns or post-processing steps in your code.

cURL

Copycurl https://api.dolphinseriesai.com/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "List the three key features of the current API tier."}]
  }'

Python

Copyfrom openai import OpenAI

client = OpenAI(base_url="https://api.dolphinseriesai.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)

Node.js

Copyimport OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.dolphinseriesai.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);

Streaming

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)