Translate lab report PDFs for patients with Google Drive, OpenAI and Gmail

Quick Overview This workflow watches a Google Drive folder for new lab-report PDFs, extracts their text, uses OpenAI (gpt-4o-mini) to structure and simplify results, translates the summary into the patient’s preferred language, and emails a Markdown report via Gmail, with error notifications sent to Slack.

How it works Triggers every minute when a new PDF is created in a specific Google Drive folder. Downloads the PDF from Google Drive and extracts the raw text content. Sends the extracted text to OpenAI (gpt-4o-mini) to extract structured patient/report fields, detect the report language, and capture raw findings. Cleans and parses the OpenAI JSON response to reliably carry forward patient details, report metadata, and findings. Uses OpenAI (gpt-4o-mini) to rewrite the raw findings into a warm, jargon-free plain-English Markdown summary with a “What to do next” section. Uses OpenAI (gpt-4o-mini) to translate the Markdown summary into the patient’s preferred language while preserving formatting. Builds a final Markdown report, converts it into a .md attachment, and sends it to the patient via Gmail. Posts error details to a Slack channel if extraction, simplification, or translation fails.

Setup Connect Google Drive OAuth2 credentials and set the folder ID in the Google Drive trigger to the lab-reports folder you want to monitor. Add an OpenAI API key credential (the workflow uses the gpt-4o-mini model for extraction, summarization, and translation). Connect Gmail OAuth2 credentials and ensure the recipient expression resolves to a real patient email (the workflow uses {{$json.patient_email}}). Connect Slack OAuth2 credentials and replace YOUR_SLACK_CHANNEL_ID with the channel where error alerts should be posted. Upload a sample PDF to the watched Google Drive folder and verify you receive the email with the Markdown attachment and any errors appear in Slack.