← Tailscale
David Ondrej
Tailscale, Clearly Explained (Beginner's Guide)
⏱ 28 min video · 3 min read13 Jul 2026Worth watching
TL;DR
David Ondrej walks through what Tailscale is, why it matters for running multiple AI agents across machines, and does a live demo installing Tailscale on a Hostinger VPS, locking it down, and deploying Hermes Agent with API keys managed via Tailscale Aperture — all orchestrated by Codex CLI.
Key points
1
Tailscale creates a private encrypted network connecting all your machines so AI agents can communicate without any open ports exposed to the internet.
2
With Tailscale, a local agent (e.g. on your MacBook) can remotely control, fix, or update agents on any VPS without manual SSH — solving the biggest pain point of multi-agent setups.
3
Tailscale Aperture is an AI gateway node on your tailnet that centrally stores API keys and secrets, so you never have to paste the same key into multiple agents on multiple machines.
4
Access control lists (ACLs) let you restrict which devices or agents can talk to which, so a compromised node cannot reach your entire network.
5
The free Tailscale tier supports unlimited devices and up to 6 users, meaning the entire networking layer costs nothing for solo builders or small teams.
Actionable insights
→
Install Tailscale on every machine and VPS you own, then immediately disable all public inbound ports and allow only Tailscale traffic — one Codex prompt can do this for you automatically.
→
Use Tailscale Aperture to store all API keys (OpenAI, Anthropic, OpenRouter, etc.) in one place so agents across different VPSs pull keys privately over the tailnet instead of storing them locally.
→
Delegate the entire Tailscale setup, VPS hardening, and agent installation to Codex (or Claude Code) by giving it high-level intent and your Cmax terminal workspace — you rarely need to type commands yourself.
→
Use Tailscale machine names instead of IP addresses when connecting agents to each other — names persist even if IPs change and make your network far easier to manage.
→
Lock in a 24-month VPS plan now (Hostinger KVM2 cited as sufficient for 20+ agents) because hardware rental prices are rising and longer terms carry the highest discounts.
Notable quotes
“99.9% of software will be used by agents. That includes tools, that includes APIs, that includes SaaS. It's going to be used by AI agents. We humans, we're just going to manage them.”
“I think 99% of slop is because of the humans. The humans don't know what to build. The humans using the AI have no taste.”
“All it takes is sending one prompt to your personal agent and it can update all of your agents to the latest and greatest model by itself — and you're saving 20, 30, 40 minutes easily.”
Worth watching?
✅
Worth watching the full video?
Watch if you are running AI agents on multiple machines and want a practical, step-by-step live demo of securing and networking them with Tailscale — the key commands, concepts, and agent-delegation workflow are all here in the summary, but the screen recording adds useful visual context for the Aperture and Cmax steps.
Saved you some time? The creator still deserves a like.
Found this by searching? That's the hard way.
summree watches the AI channels you follow and sends you the summary the moment a video drops. Nothing gets past you. From £4/month.
Try it free