Load Script ⟶

Setting the background parameter to True is recommended as deep reports can take several minutes to generate. 2. Open Source Deep Research (LangChain)

If you have access to the , you can use the o3-deep-research or o4-mini-deep-research models. This requires at least one data source, such as web search. load script

Add your API keys (OpenAI, Tavily, etc.) to a .env file. Setting the background parameter to True is recommended

Reports across multiple topics can be automated using Google Apps Script with Gemini. This requires at least one data source, such as web search

git clone https://github.com/langchain-ai/open_deep_research.git cd open_deep_research pip install -r requirements.txt Use code with caution. Copied to clipboard

import openai client = openai.OpenAI(api_key="your-api-key") # The model handles the multi-step research loop internally response = client.responses.create( model="o3-deep-research", messages=[ {"role": "user", "content": "Generate a deep report on the future of renewable energy in 2030."} ], tools=[{"type": "web_search_preview"}] # Required for deep research ) print(response.output_text) Use code with caution. Copied to clipboard

Use the provided CLI or Python entry point to start a research loop that typically runs 5–7 concurrent search iterations. 3. Bulk Deep Research (Google Apps Script)