Gemini Moderation POST Webhook for n8n: PII Redaction

Moderate user text in real time with a Gemini POST webhook—PII redaction included

This n8n automation provides a POST webhook that sends text to Google Gemini for safety moderation. You get a structured decision (allow/flag/block), toxicity level, matched categories, detected PII types, a short reason, and a cleaned version where slurs are masked and PII is redacted.

What this workflow does

• Receives a request: Handles a POST call on a webhook endpoint with JSON containing text and optional context.
• Calls Google Gemini: Sends the input to Gemini with instructions to classify safety, rate toxicity, identify categories, detect PII, and generate a redacted/masked cleaned_text.
• Parses the result: Converts Gemini’s response into a structured JSON object containing the moderation fields.
• Returns moderation output: Responds to the caller with the moderation decision and redaction results.
• Fail-safe handling: If Gemini can’t classify the content, the workflow returns a fail-safe JSON response that flags the content for human review.

Use cases

• SaaS moderation for user-submitted comments, feedback, or messages.
• PII protection when logs, tickets, or chats may contain personal data that must be redacted before storage.
• Human-in-the-loop review when confidence is low or classification fails.
• Content safety enforcement with actionable outcomes (allow/flag/block) and category tagging.

Technical details (n8n workflow)

• Webhook trigger with a /moderate endpoint (POST JSON: { "text": "...", "context": "..." }).
• Google Gemini (PaLM) API via an n8n Google Gemini Chat Model credential.
• Workflow logic using nodes such as if, code, aggregate, slack, and a sticky note.

Set up your Gemini credential, activate the workflow, then use the production webhook URL in your app to moderate content with PII redaction and safety decisions.