Developer Documentation

Base URL and Authentication

Use https://api.caiapi.cc/v1 as your base URL. Authenticate by passing your unique API key in the Authorization header. This standard approach lets you swap in any OpenAI-compatible SDK without rewriting your connection logic.

Copycurl https://api.caiapi.cc/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "Get user_id 99 for account settings."}]
  }'

Your key grants access to the uncensored model. Keep it secure, as it tracks your token usage and billing directly.

First Request with Python

Initialize the client with your key and the correct base URL. Send a simple prompt to verify connectivity and response quality.

Copyfrom openai import OpenAI

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

Expect a direct text response. The model ignores standard content filters, making it ideal for roleplay or creative writing.

Node SDK Integration

Install the official SDK and configure it to point to our endpoint. This ensures compatibility with existing TypeScript or JavaScript projects.

Copyimport OpenAI from "openai";

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

Handle the response object to extract the generated text. The structure matches standard OpenAI outputs.

Streaming Responses

Enable streaming by setting stream: true. Receive tokens as they generate, which reduces perceived latency for long outputs.

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)

Token usage details appear in the final chunk. This method is optimal for real-time chat interfaces.

Function Calling and JSON Mode

Define tools in your request to get structured outputs. Use response_format: {"type": "json_object"} to enforce strict JSON formatting.

This feature allows your application to parse model outputs programmatically. It works reliably with the uncensored model for both creative and logical tasks.

Rate Limits and Errors

Limit: 300 requests per minute. Concurrent: 8 requests per key. Error 401 means invalid key. Error 402 means no credit. Error 429 means rate limit exceeded. Context window: 100,000 tokens. Max output: 16,000 tokens.

Errors do not consume credit. Adjust your retry logic accordingly.