Guides & tutorials
Practical, production-tested guides for hosting AI agents and apps on a VPS. From securing a fresh server to deploying a self-improving agent in minutes.
Why your GPU container breaks: CUDA, drivers and pinned tags
The version mismatches that make cloud GPUs miserable — CUDA against driver, PyTorch against CUDA, and a tag that moved under you — plus what a rented GPU container genuinely cannot do.
The cheapest VPS for AI agents — what you actually need
We measured four AI agents on real servers to find the floor: 1 vCPU and 1 GB runs most of them. What that means for picking a cheap VPS, and why the server is not the bill to optimise.
Which model should you run in your coding agent?
Your agent decides which models you can use before your preference does. Provider constraints for nine coding agents, which run local models, and why the model bill dwarfs the server bill.
Self-hosted Claude Code alternatives
Open-source coding agents you can run on your own server and point at any model — including when the right answer is to keep Claude Code and just move it.
Best open-source AI coding agents (2026)
Ten open-source coding agents compared on licence, language, model independence and MCP support — plus the memory and CPU four of them actually need, measured on a real server.
Codex is four different things — which one do you want?
Codex now means a ChatGPT feature, an open-source CLI, a desktop app and an IDE extension. They are not the same product — which one runs on your own server, and which one locks you in.
Gemini CLI was retired — what to use instead
Google shut off Gemini CLI on 18 June 2026 and replaced it with the closed-source Antigravity CLI. The open-source terminal agents still worth running, with the memory and CPU each one actually needs.
Deploy your LangChain agent to a VPS
Get an agent you built with LangChain running 24/7 on your own server: clone it, keep secrets out of your code, run it as a service that survives reboots, and put HTTPS in front of it.
Run Claude Code & Codex on a cloud VPS
Spin up a remote dev box and run Anthropic's Claude Code and OpenAI's Codex agentic coding CLIs from anywhere.
How to self-host Hermes Agent on a VPS
Step-by-step guide to running Nous Research's Hermes agent on your own Ubuntu or Debian VPS — install, configure an LLM provider, set up Telegram, and run as a systemd service.
How to self-host OpenClaw on a VPS
Install OpenClaw on a Linux VPS, add your BYOK API key, connect channels (WhatsApp, Telegram, Slack, Discord), and put it behind HTTPS with Caddy.
Managed vs self-hosted AI agents: how to choose
Compare full control vs zero maintenance. A practical trade-offs table covering time, cost, updates, backups, uptime, and vendor lock-in — so you can pick the right model.
How to host WordPress on a cloud VPS
Pick a VPS, install a LEMP stack (nginx, MariaDB, PHP-FPM), configure a server block, add free TLS via Certbot, and harden your setup.
Self-host n8n for workflow automation
Run n8n with Docker, persist your data with a named volume, put it behind a reverse proxy with HTTPS, enable basic auth, and keep it updated — in under 30 minutes.
Secure your VPS: firewall, SSH keys & fail2ban
Lock down a fresh Ubuntu/Debian VPS: create a sudo user, harden SSH, configure UFW, install fail2ban, enable unattended security upgrades, and set up basic monitoring.
Skip the setup
Launch a managed AI agent or cloud VPS in about 30 seconds — no Linux expertise required. HTTPS, backups, and updates handled for you.
View managed plans