Solutions

What are you building?

Concrete specification recommendations for 31 common workloads, with the location advice that usually matters more than the hardware.

Short answer

This index covers 31 common VPS workloads, each with a concrete specification floor, a recommended plan and location, a setup path and an FAQ. The recurring lesson across all of them: for privacy-oriented workloads the jurisdiction decides more than the hardware, and for latency-sensitive ones the region decides more than the core count.

Personal VPN server

A personal VPN needs almost no CPU and almost no disk — it needs bandwidth and a jurisdiction you trust. One shared co…

1 vCPU · 1 GB · $4/month

Game server hosting

Game servers are single-thread bound, so clock speed beats core count. For Minecraft with 10–20 players, 4 GB of RAM o…

4 vCPU · 8 GB · $49/month

Forex & algorithmic trading

A forex VPS is judged on two numbers: latency to your broker's matching engine and uptime. Two dedicated cores and 4 G…

4 vCPU · 8 GB · $49/month

Crypto trading bots

Crypto exchanges concentrate their matching engines in AWS Tokyo (Binance), AWS Ashburn (Coinbase) and a handful of Eu…

4 vCPU · 8 GB · $49/month

Web scraping & data collection

Scraping is memory-bound, not CPU-bound: a headless Chromium tab costs roughly 300 MB, so plan around concurrency. Fou…

4 vCPU · 8 GB · $34/month

Seedbox & private torrent storage

A seedbox needs disk and uplink, not cores. Two shared cores, 4 GB of RAM and 2 TB of storage on a 10 Gbit port will s…

2 vCPU · 4 GB · $22/month

Self-hosted mail server

Self-hosting mail needs three things most providers do not give you: outbound port 25 open, self-service reverse DNS, …

4 vCPU · 8 GB · $34/month

Tor relay & bridge

A Tor middle relay runs comfortably on one shared core and 1 GB of RAM; the constraint is bandwidth and the provider's…

1 vCPU · 2 GB · $7/month

Docker & container host

Docker needs a real kernel, which means KVM rather than an OpenVZ or LXC container. Two dedicated cores and 4 GB run a…

4 vCPU · 8 GB · $34/month

Kubernetes nodes

A K3s control plane node needs 2 cores and 4 GB; a full kubeadm control plane wants 4 GB minimum and 8 GB to be comfor…

4 vCPU · 8 GB · $34/month

CI/CD runners

Self-hosted runners pay for themselves quickly: a Cutter 16 at $62 a month replaces several thousand hosted CI minutes…

6 vCPU · 16 GB · $62/month

WordPress & CMS hosting

A WordPress site with moderate traffic runs comfortably on two dedicated cores and 4 GB of RAM with PHP 8.3, OPcache a…

2 vCPU · 4 GB · $19/month

Proxy servers

A proxy server is bandwidth with a small CPU tax. One shared core and 1 GB of RAM saturate a gigabit port with Dante o…

1 vCPU · 1 GB · $4/month

Private DNS & ad blocking

Recursive DNS is the lightest useful workload there is: one shared core and 1 GB of RAM serve a household or a small t…

1 vCPU · 1 GB · $4/month

Nextcloud & private file sync

Nextcloud is IO-bound and storage-hungry rather than CPU-hungry. Two cores and 4 GB serve a family; a 20-person team w…

4 vCPU · 8 GB · $34/month

Matrix & XMPP servers

Synapse is memory-hungry: a small homeserver federating with busy rooms needs 4 GB, and 8 GB if you join large public …

4 vCPU · 8 GB · $34/month

Game hosting panels

Pterodactyl splits into a panel and one or more Wings nodes. The panel is light — two cores and 2 GB. Wings nodes carr…

2 vCPU · 4 GB · $19/month

Reseller & white-label hosting

Reselling needs headroom and predictability rather than raw speed. A Cutter 32 or Clipper 32 hosts a control panel, a …

8 vCPU · 32 GB · $118/month

AI & LLM inference

CPU inference is viable for small models: an 8-billion-parameter model quantised to 4 bits runs at conversational spee…

8 vCPU · 16 GB · $92/month

Media servers

Direct play needs almost nothing; transcoding needs a great deal. A Jellyfin server that only direct-plays runs on two…

2 vCPU · 4 GB · $22/month

Blockchain & validator nodes

Blockchain nodes are storage and IO monsters. A Bitcoin full node needs about 700 GB and grows; an Ethereum archive no…

6 vCPU · 16 GB · $62/month

E-commerce hosting

Shops cannot cache the pages that matter. Cart, checkout and account pages are dynamic by definition, so e-commerce is…

6 vCPU · 16 GB · $62/month

Remote development

A remote development box turns a thin laptop into a workstation. Four dedicated cores and 8 GB handle most language se…

4 vCPU · 8 GB · $34/month

Monitoring & observability

A monitoring stack is memory and disk bound. Prometheus with 15-second scrapes across 50 targets and 90 days of retent…

4 vCPU · 8 GB · $34/month

Journalism & source protection

Source-protection infrastructure is a legal architecture problem before it is a technical one. What matters is that th…

4 vCPU · 8 GB · $59/month

High availability & failover

Real availability comes from independent failure domains, not from a bigger server. Three instances in three countries…

4 vCPU · 8 GB · $34/month

DDoS-protected hosting

DDoS protection has to be always-on and upstream of your instance to be useful — mitigation that activates after an at…

4 vCPU · 8 GB · $34/month

Windows VPS

Windows Server needs materially more baseline resources than Linux: 4 GB of RAM is the practical floor and 8 GB is com…

4 vCPU · 8 GB · $34/month

Remote desktop & cloud PC

A usable remote desktop needs four dedicated cores and 8 GB of RAM — a browser with a dozen tabs will use most of that…

4 vCPU · 8 GB · $34/month

SEO tools & rank tracking

SEO tooling is memory-bound and IP-sensitive. Screaming Frog crawling 500,000 URLs wants 16 GB; a rank tracker wants s…

6 vCPU · 16 GB · $62/month

Low-latency APIs

Sub-10 ms API responses are a placement problem before they are a code problem. Physics puts a floor of about 1 ms per…

4 vCPU · 8 GB · $49/month