Xiaohongshu's Dots Studio has open-sourced dots3-note Preview and launched its official API. Supporting native ultra-long context windows and multimodal inputs, this model is currently available with limited-time free API access—making it an ideal time to test its long-context and multimodal capabilities.

Key Specifications

  • Total parameters: ~280B; Activated parameters during inference: ~16B
  • Native 512K token context window
  • Multimodal input support: text, image, video, and audio (text output)
  • Optimized for complex reasoning and long-horizon Agent workflows
Dots Studio
Dots Studio

Free Quota & Rate Limits

Rate limits are tracked per individual API Key:

  • RPM (Requests Per Minute): 30
  • TPM (Tokens Per Minute): 750,000

(Limits are subject to adjustment during the preview phase; refer to the official platform for real-time status.)

Getting an API Key

  1. Visit https://dots.ai/platform/apikeys
  2. Navigate to "API Keys" in the left sidebar to generate your key.
  3. The full key is displayed only once upon creation. Save it immediately in a secure place; if lost, you will need to generate a new one.

API Endpoint Details

  • Base URL: https://note3-prev-api.askdiandian.com
  • Model Name: dots3-note-prev
  • Authentication: Request header api-key: YOUR_KEY (Note: It does not use the standard Authorization: Bearer header)

Both OpenAI Chat Completions and Anthropic Messages protocols are officially supported.

Client Configuration Guide

Most "OpenAI-compatible" clients can integrate seamlessly. However, since the authentication header requires api-key rather than the standard Bearer token, certain tools may require custom header configuration. Below are the most common integration methods.

1. Verify via cURL

curl -X POST https://note3-prev-api.askdiandian.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "api-key: YOUR_API_KEY" \
  -d '{
    "model": "dots3-note-prev",
    "messages": [
      {"role": "user", "content": "Hello! Please introduce yourself briefly."}
    ],
    "stream": false,
    "max_tokens": 512
  }'
Successful cURL Response
Successful cURL Response

Once you verify a valid response, proceed with configuring your third-party clients.

2. Python (OpenAI SDK)

from openai import OpenAI
 
client = OpenAI(
    api_key="placeholder",  # Bypasses standard Authorization
    base_url="https://note3-prev-api.askdiandian.com/v1",
    default_headers={"api-key": "YOUR_ACTUAL_API_KEY"}
)
 
response = client.chat.completions.create(
    model="dots3-note-prev",
    messages=[{"role": "user", "content": "Hello"}],
    max_tokens=1024
)
print(response.choices[0].message.content)

3. Cherry Studio

  1. Open "Settings → Model Providers".
  2. Add a custom provider and select "OpenAI Compatible".
  3. Set the API Base URL to: https://note3-prev-api.askdiandian.com
  4. Enter your API Key in the API Key field.
  5. Manually add the model name: dots3-note-prev
Configuring Xiaohongshu LLM in Cherry Studio
Configuring Xiaohongshu LLM in Cherry Studio

Best Practices & Tips

  • As this is a preview release, performance and stability are actively iterating and may vary depending on task complexity.
  • Ideal for exploring 512K long-context and multimodal tasks; not recommended for mission-critical production environments yet.
  • Keep your API Key secure—never commit it to public repositories or expose it in frontend builds.
  • If you encounter a 401 Unauthorized error, verify first that your custom header is set to api-key.
  • Multimodal inputs (images/videos/audio) follow standard official formatting conventions.

Overall, onboarding is straightforward. The non-standard auth header is the primary catch to watch out for. Running a quick cURL verification first will save troubleshooting time later.

Vibe Coding