← Back to VecRoute

Privacy Policy

Last updated: October 2026

VecRoute Pte. Ltd. ("VecRoute", "we", "us") is dedicated to zero-data retention routing. This Privacy Policy outlines how our semantic router and cache-aware AI gateway handle data when you route inference queries through our infrastructure.

Key Guarantee: VecRoute operates on an in-memory, zero-persistence architecture. We never store, log, or train on your prompt tokens, tool outputs, thinking tokens, or model completions.

1. Information We Collect

We process only the telemetry required for rate limiting, billing, and routing latency optimization:

2. Data Processing & In-Memory Architecture

When your application or Model Context Protocol (MCP) client dispatches a query through VecRoute:

3. Third-Party Model Providers & MCP Servers

Payloads are transmitted strictly to upstream model providers (e.g., Anthropic, OpenAI, Google Cloud Vertex, DeepSeek) or customer-configured MCP servers as dictated by your routing rules. Upstream processing is governed by each provider's enterprise API terms (which guarantee zero training on API data).

4. Regulatory Compliance (EU AI Act, GDPR & Singapore PDPA)

VecRoute complies with the EU AI Act (2026 regulatory framework) regarding general-purpose AI transparency, GDPR, and the Singapore Personal Data Protection Act (PDPA). You retain the right to audit, export, or purge all organizational metadata by contacting privacy@vecroute.xyz.

5. Security Controls

All traffic is protected by TLS 1.3 in transit. Gateway keys are cryptographically hashed using Argon2id. Infrastructure runs on SOC 2 Type II certified edge data centers with strict role-based access control and zero persistent disk logging for request bodies.

6. Contact Us

For data privacy queries or compliance verification, contact:
privacy@vecroute.xyz
VecRoute Pte. Ltd., Singapore