AI Terminal

Security-first AI gateway notes · Cloudflare edge deployment · practical engineering logs

This site runs a Cloudflare Worker based multi-model AI gateway with Grok / GPT / Codex switching, SSE streaming, and layered abuse protection.

Featured

• komari from Zero to Usable: A Beginner-Friendly Guide to Getting It Installed and Actually Useful

  Published: 25 Mar, 2026

  | 📝 Updated: 25 Mar, 2026

  A practical guide for complete beginners to Komari: how to choose a deployment method, what to check after your first login, how to add nodes, switch themes, enable offline/expiry/load alerts, run latency tests, and avoid the most common pitfalls.

• Don’t Force PVE onto Oracle ARM: How I Split One Host into Two Maintainable Mini VPSes with Incus Routed

  Published: 25 Mar, 2026

  When nested KVM isn’t available on Oracle ARM, forcing PVE usually creates more complexity than value. In this setup, I used a Debian host plus Incus routed containers to split one machine into two lighter, more stable, and much easier-to-maintain mini VPS-style nodes.

• OpenClaw Memory Archiving, Compression, Reset, and Recovery Workflow

  Published: 9 Mar, 2026

  A direct explanation of OpenClaw’s current workflow for long-term memory archiving, compression, reset preparation, and post-reset recovery.

• From a Web Mirror to an AI Assistant on My Phone: I Built an Android App in Two Days (Vibe Coding Practice Log)

  Published: 4 Mar, 2026

  A two-day Android Aura build log: moving from web mirror sessions to persistent mobile AI assistant experience with multi-model collaboration and release-stage stability fixes.

• How I Built a Secure Multi-Model AI Gateway with Cloudflare Workers (Grok + GPT/Codex)

  Published: 2 Mar, 2026

  A practical architecture log of building a security-first AI gateway on Cloudflare Workers + KV, with SSE streaming passthrough, dual-layer rate limiting, and prompt hardening.