This is an n8n community node that integrates Akto into your n8n workflows. It lets you validate prompts with Akto Guardrails and ingest LLM interactions for API security monitoring.
- Guardrails — Validate a user prompt before it reaches your AI Agent. Blocked prompts are routed to a separate output; allowed prompts continue normally.
- Ingestion — Ingest the user prompt and AI Agent response into Akto for API security monitoring and observability.
You need an Akto API credential with:
- Akto Data Ingestion URL — Base URL of your Akto data ingestion service (e.g.
http://localhost:8080) - Akto API Token — API token sent as the
Authorizationheader on every request - Timeout (seconds) — Request timeout (default: 5s)
Tested with n8n v1.x.
- Add the Akto node before your AI Agent with operation set to Guardrails
- Connect the Allowed output to your AI Agent
- Add another Akto node after the AI Agent with operation set to Ingestion
- The Ingestion node references the Guardrails node output by default — update the Prompt field if you rename the Guardrails node
The Guardrails node fails open: if the Akto service is unreachable, prompts are allowed through automatically.