n8n Drive-OpenAI Workflow

I’m looking for a clean, no-code n8n configuration that ties Google Drive, OpenAI, and Google Sheets together. At the moment I don’t have n8n running anywhere, so the very first step is helping me spin up a fresh instance (Docker on a small VPS is fine, but I’m open to whatever method you prefer). All Google service and OpenAI credentials are already in hand; I simply need them wired in securely.

Workflow logic I need: when I trigger the workflow on demand, it should scan a specific Google Drive folder for PDFs that have not been processed before, perform OCR if necessary, pull out all text, and send that text to OpenAI. The AI should return a structured payload that contains a short summary, a detailed summary, key points, keywords, named entities, relevant quotes, and inline citations. n8n then records that payload—together with basic file metadata such as file name, link, size, and last modified date—into a designated Google Sheet. If the same file appears unchanged on a later run, the workflow must recognise it and skip reprocessing unless the file has been updated.

Please stick to native n8n nodes and community-maintained nodes only, no custom JavaScript functions or external micro-services. I’d also like the nodes named clearly, credentials stored in the recommended way, and the whole thing commented so I can adjust prompts or add new columns later.

Deliverables
• A working n8n workflow (exported .json or installed directly) meeting the above logic
• A short, plain-language maintenance guide covering installation, credential setup, and how to trigger or edit the flow

I’ll test the setup by uploading a new PDF and confirming the Sheet populates correctly with no duplicates. Once that passes, the project is complete.