Skip to main content

4 posts tagged with "Security"

Data isolation, audit trails, and compliance posture.

View All Tags

OpenProxyAI vs. Portkey: SaaS-first vs. self-hosted-first

· 2 min read
Founder, OpenProxyAI

Portkey is the most feature-comparable product to OpenProxyAI in this space — a unified gateway, real guardrails (including PII redaction), observability, and enterprise governance. If you put the feature lists side by side, they overlap a lot. The difference that actually matters is architectural, not a checklist item: which deployment model is the default, and which one you have to pay to unlock.

Compliance templates: policy configs, not badges

· 3 min read
Founder, OpenProxyAI

We shipped three compliance templates this phase — Healthcare, Finance, and Government — and the most important thing to say about them is what they're not: they're not a claim that OpenProxyAI holds any certification on your behalf. They're pre-built policy configurations you apply, review, and enforce yourself.

Row-level security: making data leaks impossible by construction

· 3 min read
Founder, OpenProxyAI

Every multi-tenant system has to answer the same question: what stops a bug in organization A's code path from returning organization B's data? The conventional answer is discipline — every query gets a WHERE org_id = ? clause, added by convention, reviewed in code review, hopefully never forgotten.

We didn't want the answer to be discipline. We wanted it to be a database property that doesn't care whether the application code remembered.

Why OpenProxyAI runs in your VPC, not ours

· 2 min read
Founder, OpenProxyAI

Most AI gateways you'll evaluate are built SaaS-first: a multi-tenant service you send your traffic to, with a self-hosted or "VPC" option available if you pay for the enterprise tier. OpenProxyAI is built the other way around. Self-hosted is the default, not the upsell.

That ordering isn't a licensing decision. It changes what you're actually agreeing to when you adopt the tool.