deepseek 0.0.1.dev3

Deepseek API Library

DeepSeek API Library

DeepSeek-V3 delivers groundbreaking improvements in inference speed compared to earlier models. It leads the performance charts among open-source models and competes closely with the most advanced proprietary models available globally.

This Python library provides a lightweight client for seamless communication with the DeepSeek server.

For deepseek GUI support, welcome to check out DeskPai.

---

📥 Installation

pip install deepseek

---

🥔 Preparation - DEEPSEEK_API_KEY

You need to obtain a DeepSeek API Key. If you don't have one, visit here to generate it.

You can configure your API key as an environment variable.

On macOS or Linux:

export DEEPSEEK_API_KEY=<YOUR_API_KEY>

On Windows (PowerShell):

setx DEEPSEEK_API_KEY <YOUR_API_KEY>

If DEEPSEEK_API_KEY is not set, you need to manually pass api_key in code DeepSeekAPI(api_key).

---

💎 Usage

Initialize the API Client

from deepseek import DeepSeekAPI
api_client = DeepSeekAPI()

Retrieve Account Balance

api_client.user_balance()

List Available Models

api_client.get_models()

Chat (Streaming Disabled)

response = api_client.chat_completion(prompt='Hi')
print(response)

Chat (Streaming Enabled)

for chunk in api_client.chat_completion(prompt='Hi', stream=True):
    print(chunk, end='', flush=True)

Chat (Multi-turn Mode)

For multi-turn mode, you need to construct prompt as a list with chat history. An example is as below:

prompt = [
    {"role": "system", "content": "You are a helpful assistant"},
    {"role": "user", "content": "What is the capital of China?"},
    {"role": "assistant", "content": "The capital of China is Beijing."},
    {"role": "user", "content": "What is the capital of the United States?"}
]
for chunk in api_client.chat_completion(prompt=prompt, stream=True):
    print(chunk, end='', flush=True)

This is another multi-turn chat example in Deskpai Image Chat.

Fill-In-the-Middle (Streaming Disabled)

response = api_client.fim_completion(prompt='Hi', max_tokens=64)
print(response)

Fill-In-the-Middle (Streaming Enabled)

for chunk in api_client.fim_completion(prompt='Once upon a time, ', stream=True):
    print(chunk, end='', flush=True)

Customized Model Inference Parameters

use_case = 'Creative Writing'
kwargs = {'max_tokens': 7680, 'temperature': TEMPERATURE_MAP[use_case]}
api_client.chat_completion(prompt='Hi', **kwargs)

---

🔗 Contact

Maintained by deskpai.com 2025
Contact: [e-mail]