deep-research

Automate deep research by running a Python script that sends your prompt to the OpenAI Deep Research API, which gathers web evidence and outputs a Markdown report with cited sources. The process runs synchronously to produce a finished, reproducible analysis.

Yes, you need an OpenAI API key to run reproducible web search research. You also need Python 3.7 or higher and active internet access for the API to gather online evidence and generate the timestamped Markdown report.

Prompt enhancement for in-depth topic analysis detects vague inputs and guides you through 2 to 3 clarifying questions. This refines your query before execution, ensuring the web-enabled research returns focused, relevant evidence and minimizes token waste.

Yes, you can use an optional prompt-file input instead of passing a direct string for deep research execution. The script supports flexible configuration, allowing you to load prompts from a file and set custom model and timeout options.

The output of an automated deep research report includes a complete Markdown document with a dedicated sources section and a metadata footer. It also saves your enhanced prompts with timestamps locally to ensure full reproducibility.

Synchronous web-enabled research execution runs in a single long-running flow to minimize token waste, meaning the script blocks until the API completes. You must account for this wait time and configure appropriate timeout options for complex topic analysis.