How to Call DeepSeek-R1 API with Python: A Step-by-Step Guide

DeepSeek's R1 large language model offers powerful capabilities for various applications. This tutorial provides a detailed, beginner-friendly guide on how to call the DeepSeek-R1 API using Python. Whether you're new to programming or an experienced developer, this guide will walk you through the process with clear instructions and visuals.

Preparing for API Call

1. Get an API Key

First, you need an API key to access DeepSeek's services.

1. Go to the DeepSeek console and log in.
2. New users sometimes receive a complimentary token credit (check if available; otherwise, you'll need to purchase tokens).
3. Navigate to the "API Keys" section and click "Create API Key."
4. Copy the generated key (e.g., sk-123456789abc) and store it securely. You'll need it later.

Python Calling Steps

1. Environment Preparation: Install Python

If you haven't already, install Python on your computer. macOS comes with Python 3 pre-installed. You can download the latest version from python.org. It's recommended to download version 3.8 or later.

Important: Make sure to check the "Add Python to PATH" option during installation.

2. Install the Requests Library

The requests library simplifies making HTTP requests in Python. Open your IDE tool (like VSCode) and run the following command in the terminal (Terminal → New Terminal):

pip install requests
# or
pip3 install requests

A successful installation will show a message like "Successfully installed requests-x.x.x."

Troubleshooting: If the installation fails, especially with multiple Python versions on your system, ensure you're using the correct pip version for your desired Python environment. VSCode's "Python" extension can help manage environments and terminal selection.

3. Basic Calling Code

Create a new Python file (e.g., deepseek.py) and paste the following code, replacing "sk-your_key" with your actual DeepSeek API key:

import requests

# Replace with your actual API Key
API_KEY = "sk-your_key"
url = "https://api.deepseek.com/chat/completions"

headers = {
    "Content-Type": "application/json",
    "Authorization": f"Bearer {API_KEY}"
}

data = {
    "model": "deepseek-reasoner",  # Use 'deepseek-reasoner' for R1 or 'deepseek-chat' for V3
    "messages": [
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Who are you?"}
    ],
    "stream": False  # Turn off stream
}

response = requests.post(url, headers=headers, json=data)

if response.status_code == 200:
    result = response.json()
    print(result['choices'][0]['message']['content'])
else:
    print("Request failed, error code:", response.status_code)

Key Points:

• model: Specifies the DeepSeek model to use. Use "deepseek-reasoner" for R1 and "deepseek-chat" for V3.

4. Run the Code

Execute the script in one of these ways:

• Click the "▶" button in your IDE.
• Run python3 deepseek.py or python deepseek.py in the terminal.
• Right-click in the editor and choose "Run Python File in Terminal."

5. Success!

A successful call will output the model's response to your question. For example:

Hello! I am the DeepSeek-R1 that is developped by the DeepSeek company. I can help with any problems you have.

Code Breakdown

1. Key Parameters Explained

2. Multi-Turn Conversations

To engage in multi-turn conversations, structure the messages list as follows:

messages = [
    {"role": "system", "content": "You are a poet."},
    {"role": "user", "content": "Write a poem about spring."},
    {"role": "assistant", "content": "Spring breeze brushes long willow branches..."},
    {"role": "user", "content": "Please continue with a second verse."}
]

3. Streaming Mode

To enable streaming output, modify the following parameters:

data["stream"] = True
response = requests.post(url, headers=headers, json=data, stream=True)

for line in response.iter_lines():
    if line:
        decoded_line = line.decode('utf-8')
        print(decoded_line)

Common Issues

• Q1: How to distinguish between the V3 and R1 Model.
  • V3: model: "deepseek-chat"
  • R1: model: "deepseek-reasoner"
• Q2: 401 Error: Verify your API key is correct and hasn't expired.
• Q3: "No module named 'requests'": Ensure pip install requests was executed in the correct environment/terminal. Check the Python interpreter in VSCode.
• Q4: requests.exceptions.JSONDecodeError: Expecting value: line 1 column 1: This usually indicates a server-side issue (busy server). The API might be returning an empty response.
• Q5: Streaming output shows "keep-alive": Similar to Q4, this suggests the server is overloaded and maintaining a long connection.

Alternative: Using Apifox for API Calls

For a simpler approach, consider using Apifox to interact with the DeepSeek-R1 or V3 API with or without streaming option. For more detailed instructions, see this article: Deepseek API Debugging.

References

• Apifox
• DeepSeek console
• Python downloads

This tutorial provides a comprehensive foundation for calling the DeepSeek API. Start with basic conversations, gradually trying to implement more advanced features like streaming and multi-turn dialogues.