Quickstart: uncensored api integration

Authentication: using your claude key

Start by obtaining your API key from the dashboard. Paste it into your environment variables or client configuration. The service validates credentials against the /v1 base path. Use the Authorization: Bearer header for every request. Invalid keys return a 401 error immediately.

• Sign up with Google or email.
• Copy the key shown on the dashboard.
• Store it securely in your CI/CD pipeline.

Copycurl https://api.claudekey.cc/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "List three reasons to use custom api keys for local models"}]
  }'

Endpoint: /v1/chat/completions

Send POST requests to the chat completions endpoint. The service accepts standard message arrays with role and content fields. This is the core mechanism for text generation. You can test connectivity with a simple system prompt. The endpoint handles all routing logic internally.

• Use POST /v1/chat/completions.
• Include your model name in the body.
• Receive structured JSON responses.

Copyfrom openai import OpenAI

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

Model ID: use 'uncensored'

Specify the model ID as uncensored in your requests. This identifier routes your prompt to the correct inference server. The model is tuned for fewer refusals on adult content. It does not route to other vendors like GPT or Claude. Always verify the model name in your client config.

• Set model: "uncensored" in your payload.
• Verify the response comes from the correct server.
• Check the /v1/models endpoint for verification.

Copyimport OpenAI from "openai";

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

Streaming and token usage

Enable streaming by setting stream: true in your request. The server returns chunks via Server-Sent Events. Token usage metrics appear in the final chunk of the stream. This allows real-time UI updates for users. It reduces perceived latency significantly. Monitor the usage object carefully.

• Parse SSE events in your client.
• Extract usage data from the last chunk.
• Handle partial content gracefully.

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)

Function calling and JSON mode

Use the tools parameter for function calling. Define your schema in the request body. The model returns structured tool calls. You can also force JSON output using response_format: {type: "json_object"}. This ensures strict parsing for downstream applications. Both features work seamlessly with the uncensored model.

• Define tool schemas in your payload.
• Parse the model's tool call arguments.
• Enforce JSON mode for API integrations.

Parameters: temperature, top_p, and more

Adjust temperature for creativity or top_p for diversity. Set max_tokens to limit output length. The model supports up to 16,000 output tokens per request. The total context window is 100,000 tokens. Use seed for deterministic results. These parameters give you fine-grained control over generation.

• Set temperature between 0 and 2.
• Limit max_tokens to prevent overflow.
• Use seed for reproducible outputs.