System Design
Architectural Overview
The Spendplane Architecture is designed for Zero-Trust AI Autonomy. It provides a secure, high-speed tunnel between your LLM (v0, Cursor, Bolt) and internal VPC resources.
Data Flow: LLM to Internal VPC
LLM AGENT
SPENDPLANE
PRIVATE VPC
The Shadow Plane
Unlike traditional VPNs, Spendplane does not require a static public IP. It uses Ephemeral Shadow Planes—isolated proxies created on-the-fly that only live as long as the LLM session. Metadata is stripped, and PII is redacted before reaching the VPC through our Vibe Debugger layer.
Governance & Compliance
Every interaction is double-audited. The Spendit Agent provides real-time traffic inspection to ensure that only authorized API calls and database queries are dispatched. All logs are vault-indexed for HIPAA and internal compliance workflows.
Network Topology Status: Verified
Transit
WireGuard/v2
Auth
Mutual TLS (mTLS)
Shield
Vibe-Redact Enabled