Dolphin AI Integration Guide

Prerequisites: SDK and API key setup

Install the official SDK for your language. Create an account via email or Google login on our dashboard. Generate your API key immediately. No phone number is required. The key appears once on screen. Store it securely. We do not store your payment method. Crypto top-ups use USDT or USDC. Credit never expires. Errors cost nothing. You pay only for tokens used.

Configuring the base URL and headers

Point your client to our base URL. Set the authorization header with your key. Use the model ID "uncensored". This ensures you hit our uncensored model. It is not GPT or Claude. The endpoint accepts standard chat completions. Headers must include the API key. Other fields are optional. Temperature and top_p work as expected. We support standard OpenAI parameters. Check the documentation for limits.

Copycurl https://api.dolphinllmmodel.com/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "What is the max token limit for this endpoint?"}]
  }'

Sending your first chat completion request

Send a POST request with your messages. Include the system prompt for context. Define the user message content. The model returns a text response. You can specify max tokens. Default output is 2,048 tokens. Set higher limits if needed. The context window is 100,000 tokens. Input and output share this limit. Ensure your prompt fits. Errors return standard JSON codes. Check the response body for errors.

Copyfrom openai import OpenAI

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

Enabling streaming for real-time responses

Set streaming to true in your request. The API returns chunks via SSE. Each chunk contains partial text. Process tokens as they arrive. This reduces perceived latency. You get immediate feedback. The final chunk includes token usage. Monitor usage for cost tracking. Streaming works with all parameters. Tools and JSON mode also stream. Handle connection drops gracefully. Reconnect if the stream fails.

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)

Using tools and function calling

Define function schemas in your request. Pass them as tools to the API. The model returns tool calls. Parse the arguments for execution. Re-submit results to the model. This enables complex workflows. Support is standard OpenAI format. Tool choice can be automatic. You can force specific tools. JSON mode works with tools too. Ensure your schema is valid. Invalid schemas cause errors. Test with simple functions first.

Copyimport OpenAI from "openai";

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

Handling rate limits and errors

We enforce 300 requests per minute. Concurrent requests are limited to eight. Request bodies must be under 8 MB. A 401 error means an invalid key. Check your authentication header. A 402 error indicates no credit. Top up with crypto to continue. 429 means you hit the rate limit. Wait and retry with backoff. Errors are free. Only completed tokens cost credit. Context is 100,000 tokens total.