Skip to content

README.md

Latest commit

 

History

History
149 lines (95 loc) · 4.09 KB

README.md

File metadata and controls

149 lines (95 loc) · 4.09 KB

🔬 Deep Research with OpenAI API

中文文档 | English

Build your own deep research tool powered by the OpenAI API. This Python script automates the research process: it first uses GPT-4o to interactively clarify your research question, then refines it into a comprehensive prompt, and finally executes the deep research to generate a detailed report.

🤷 Why Do We Need This Tool?

ChatGPT Plus subscription provides only very limited deep research opportunities. With this tool, you can:

  • 🔥 Use the OpenAI API to perform unlimited deep research (pay-as-you-go)
  • 💰 Have full control over your usage and costs
  • 🛠️ Customize the research process to fit your specific needs

✨ Features

  • 🤔 Interactive Question Clarification: Uses GPT-4o to ask clarifying questions to better understand your research needs.
  • 📝 Automated Prompt Engineering: Synthesizes your original question and answers into a high-quality, detailed prompt for the research model.
  • ⚡ Dual Model Support: Choose between a faster, lightweight model (o4-mini-deep-research) or a more comprehensive one (o3-deep-research).
  • 📊 Automatic Report Generation: Calls the Deep Research API and saves the output as a timestamped Markdown file.

🚀 How to Use

1️⃣ Prerequisites

  • Python 3.x
  • An OpenAI API key.

2️⃣ Clone the Repository

git clone https://github.com/ZhishanQ/Deep-Research-with-OpenAI-API.git
cd Deep-Research-with-OpenAI-API

3️⃣ Install Dependencies

You need to install the following Python packages:

pip install openai

📦 Required packages:

  • openai - The official OpenAI Python client library (the newest version is recommended)

⚠️ If your code reports errors, please upgrade openai to newest version by:

pip install --upgrade openai

4️⃣ Set Up Your API Key

You need to set your OpenAI API key as an environment variable.

🍎 On macOS/Linux:

export OPENAI_API_KEY='your-api-key-here'

🪟 On Windows:

set OPENAI_API_KEY=your-api-key-here

💡 Note: Replace 'your-api-key-here' with your actual OpenAI API key.

📌 Permanently add to shell profile (Optional) - Click to expand

Or permanently add it to your shell profile (e.g., ~/.bashrc, ~/.zshrc) by:

🐧 Linux:

echo "export OPENAI_API_KEY='your-api-key-here'" >> ~/.bashrc
source ~/.bashrc

🍎 MacOS:

echo "export OPENAI_API_KEY='your-api-key-here'" >> ~/.zshrc
source ~/.zshrc
🌐 Using a Custom API URL (Optional) - Click to expand

If you're using a third-party API provider instead of the official OpenAI API, you need to customize the API base URL.

Method 1: Set via environment variable

export OPENAI_BASE_URL='https://your-custom-api-url.com/v1'

Method 2: Modify the code directly

Open main_en.py or main_zh.py and modify the client initialization section:

# Find this section at the beginning of the file
client = OpenAI(
    api_key=os.getenv("OPENAI_API_KEY"),
    base_url="https://your-custom-api-url.com/v1",  # Add this line
    timeout=3600
)

💡 Note: Replace https://your-custom-api-url.com/v1 with your actual API endpoint URL.

5️⃣ Configure the Script

Open main_en.py (for English interface) or main_zh.py (for Chinese interface) in a text editor. Find the following line and replace the placeholder text with your research topic:

original_question = r"""your research question goes here"""

6️⃣ Run the Script

Execute the script from your terminal:

# For the English version
python main_en.py

# For the Chinese version
python main_zh.py

The script will guide you through the process of clarifying your question and then perform the research. The final report will be saved as a .md file in the same directory.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.