Hi everyone,
While scaling automated workflows using the OpenAI node and HTTP Request nodes in n8n, two operational bottlenecks kept coming up:
1. High token consumption caused by repeated, near-identical customer queries.
2. The risk of prompt injections altering system instructions or extracting sensitive prompts.
To solve this, I built **Space Guardrail L7**, a lightweight gateway compatible with the OpenAI API spec.
### How it integrates into n8n:
It works as a drop-in replacement. You just override the `Base URL` in your OpenAI credentials or HTTP Request node:
* **Base URL:** `https://api.corenode.com.mx/v1\`
* **Auth:** Bearer Token (`sk-space-…`)
### Core Features & Benchmarks:
* **Sub-2ms Semantic In-Memory Cache:** Hits local RAM/Redis directly. Repetitive queries drop response latency from ~800ms down to 1.2ms with zero upstream token costs.
* **L7 Injection Shield:** Inspects incoming payloads before proxying upstream, mitigating delimiter confusion and system override vectors.
* **Forensic Audit Logs:** Exports structured telemetry logs in CSV/JSON format for risk tracking and client compliance.
I’m provisioning free beta access for 10 workflow builders to benchmark latency and token savings in production or staging workflows. Each key includes **50 free requests**.
Drop a reply below or send me a DM if you’d like an API key to test it out!