DuckDuckAI is a Python package for interacting with DuckDuckGo's chat API. It allows you to fetch responses from DuckDuckGo's AI models and print them in a streamed format or as a complete response.
To install the DuckDuckAI package, you can use pip:
pip install DuckDuckAIYou can interact with DuckDuckAI by calling the ask function. It supports both streaming responses or returning the entire message at once.
from duckduckai import ask
# Fetch response in streamed format (printing character by character)
ask("Tell me a joke", stream=True)[0] # Use 0 if you want to retrieve the response only (No Cookies)
# Fetch response as a complete message
response = ask("Tell me a joke", stream=False)
print(response)| Parameter | Type | Description | Default |
|---|---|---|---|
| query | str | The search query string. | Required |
| stream | bool | Whether to stream results or fetch them all at once. | True |
| model | str | The model to use for the response. | gpt-4o-mini |
DuckDuckAI currently supports the following models:
| Model ID | Description |
|---|---|
gpt-4o-mini |
A smaller variant of GPT-4o designed for quick, concise responses with less computation. |
meta-llama/Llama-3.3-70B-Instruct-Turbo |
Meta's large-scale Llama 3.3 model with 70 billion parameters designed for fast and accurate responses. |
claude-3-haiku-20240307 |
Anthropic's Claude 3 Haiku model optimized for efficient, high-quality responses. |
mistralai/Mistral-Small-24B-Instruct-2501 |
Mistral AI's 24 billion parameter model trained for instruction-based tasks. |
o3-mini |
OpenAI's compact reasoning model optimized for lightweight performance. |
Additional models may be available but subject to access restrictions. Some models may require specific permissions or may not be available in all regions.
You can reuse the authentication token to make multiple requests more efficiently:
from duckduckai import ask, fetch_x_vqd_token
# Fetch a token once
token = fetch_x_vqd_token()
# Use the same token for multiple requests
response1 = ask("What is quantum computing?", model="gpt-4o-mini", token=token)[0] # Do not put [0] if you want the token in the response
response2 = ask("Explain neural networks", model="claude-3-haiku-20240307", token=token)[0] # Do not put [0] if you want the token in the responseThis project is licensed under the Apache-2.0 license - see the LICENSE file for details.