DeepKeep icon

DeepKeep

Interact with the DeepKeep AI Firewall API

Actions3

Firewall Conversation → Check Input

AI-generated

Summary

Check a text input against the guardrails defined within a specified DeepKeep firewall conversation to detect violations.

Inputs

  • Firewall Name or ID (required) — Select the firewall that contains the guardrails to check against, either by choosing from a list or specifying an ID with an expression.
  • Conversation ID (required) — Provide the ID of the conversation within the firewall to check the input against.
  • Content (required) — Enter the text content (prompt) that you want to verify for compliance with the firewall guardrails.
  • Return Full Response (Enable Logs) — Enable to receive a detailed log of all detected violations; when disabled, only the primary violation may be included in the response.

Output shape

a single JSON object containing a 'results' property with the detected violations from the firewall check

The response includes all detected violations if logs are enabled; otherwise, it may only return the primary violation. The node handles one input item per execution and returns corresponding output per item.

Examples

Example 1: Check a user prompt against firewall guardrails

Specify the target firewall and conversation IDs, provide the input text to check, and optionally enable logs to get detailed violation information.

Links

Discussion