Ideas from an independent AI company
The private
intelligence
journal.
Practical writing about privacy, infrastructure and building useful AI without surrendering control.
The Cheapest OpenAI-Compatible API in 2026 (That Doesn't Log Your Data)
An OpenAI-compatible API with flat input/output pricing and zero prompt retention on our gateway. Here's how it compares to Groq, OpenRouter, and DeepInfra on price, speed, and privacy.
GDPR-Compliant LLM APIs: A Practical Guide for EU Builders
What 'GDPR-compliant LLM API' actually means in practice — zero retention, sub-processor clarity, and data residency — and the checklist to evaluate any provider.
Private AI Compute in Australia: Data Residency & Zero-Log Inference
Why Australian businesses and agencies are moving AI inference onshore — latency, data residency, and ESG procurement — and what private AI compute actually requires.
What Is Zero-Data-Retention AI? (And Why It Matters for Your Agents)
Prompt retention, upstream retention, and ephemeral memory — what the terms actually mean and what to verify before sending agent data.
Zero-Log AI Inference: What It Actually Means and How to Verify It
How to assess prompt-retention claims, distinguish gateway controls from upstream terms, and verify a provider before sending sensitive data.
The Security Checklist for AI Agent Backends (Before You Delegate Anything)
Agents read your email, code, and contracts — then act on your behalf. This is the backend security checklist to run before you give an agent that much power.
HIPAA and AI APIs: What Zero-Data-Retention Actually Buys You
A practical look at whether a zero-retention AI API satisfies HIPAA requirements, what a BAA can and can't cover, and how to evaluate a provider for regulated workloads.