Kamatera (complete cluster)
Kamatera review
CPU types A–D, monthly vs hourly, power-off residual and scenario hypotheses.
UpCloud
MaxIOPS prepaid families, shut-down vs delete, Simple Backup leftovers, Fair Transfer, DIY n8n Compose.
Order to verified host
Calculator selections, order-summary audit and SSH day-zero baseline.
Ollama & OpenClaw on CPU
Conditional RAM hypotheses, private binds, GPU-elsewhere exits.
Marketplace vs Compose
Image discovery kept separate from operator-owned stacks and restore proof.
Billing, backup and cancel
Export proof, terminate vs power-off, Disable Account and invoice lag.
ScalaHosting (complete cluster)
Managed WordPress/SPanel cloud versus Unmanaged root VPS. Five articles make the responsibility boundary explicit, pair intro versus renew prepaid pricing, keep official n8n packaging separate from DIY Compose, refuse most local-AI fits, and require export proof before Client Area Request Cancellation.
ScalaHosting review
Managed vs Unmanaged, intro/renew ledger, support exclusions and scenario hypotheses.
Order to verified host
Order summary, SPanel Admin first account or Unmanaged SSH discovery.
When not to run local models
Managed is rarely an AI appliance; bounded Unmanaged CPU experiments only.
Managed n8n vs DIY Compose
Official packaging path kept separate from operator-owned stacks.
Billing, backup and cancel
Export proof, Request Cancellation, renew honesty and backup conflicts.
Vultr (complete cluster)
Composable shared-to-dedicated cloud with regional product and price choices. Five articles distinguish hourly billing from a free stop, preserve VX1 recovery unknowns and the API/marketing region discrepancy, and make final-cutoff restore verification a precondition for destruction.
Vultr VPS review
Shared and dedicated CPU choices, products, regions, transfer and conditional scenario fit.
From order to verified host
Product/region selection, SSH, two-layer firewall, listener inventory and rollback timing.
Model fit and private endpoints
External-model architecture, bounded CPU inference, Ollama, OpenClaw and separate GPU paths.
n8n deployment and operations
Pinned Compose, private Postgres, Caddy HTTPS, lifecycle gates and safe updates.
Billing, backup and cancellation
Final-cutoff export/restore before destroy, independent copy and remaining-charge audit.
InterServer (complete cluster)
Monthly linear Linux and Storage slices, three U.S. regions and manual KVM recovery controls. The cluster distinguishes monthly arithmetic from annual commitments, broad support headlines from application management, and templates from production configuration.
InterServer VPS review
$3 monthly slice arithmetic, U.S. regions, product choices, support scope and scenario fit.
Template and SSH setup
Order evidence, VNC recovery, cloud-init inventory, SSH/firewall verification and DNS gates.
Slice sizing and private Ollama
Hosted-model architecture, bounded CPU tests, template naming traps and GPU/split exits.
Custom n8n Compose
Pinned n8n/Postgres/Caddy, protected secrets, internal networking, readiness and safe updates.
Billing, backup and cancellation
Monthly renewal, manual KVM/45-day backup boundaries, checked exports, Restic and cancel-last.
RackNerd (complete cluster)
Annual KVM specials and a broad ctrl one-click catalog for bounded self-hosting. The cluster distinguishes standard KVM, annual specials, Ryzen, AI-app marketing and GPU dedicated servers; it builds independent exports rather than assuming provider snapshots.
RackNerd VPS review
Annual cash arithmetic, product families, regions, ctrl, support and scenario sizing.
Safe ctrl and SSH setup
Destructive-install gate, host verification, VNC recovery and two firewall layers.
2 GB, 4 GB and 8 GB decisions
External-model architecture, bounded Ollama tests, private endpoints and split triggers.
n8n operations
Discover the delivered image, set limits, check readiness and update without wiping.
Billing, backup and cancellation
Renewal terms, SQLite/Postgres exports, Restic verification, isolated restore and cancel-old-last.
Contabo (complete cluster)
Germany-rooted VPS provider; strong RAM-per-euro story; fair-use unlimited traffic; many regions with EU as base pricing. Five long-form English articles:
Full Contabo VPS review
Company, pricing with verify labels, DCs, praise and pain, who it fits, AI angle.
Hands-on operations guide
Order → panel → SSH → firewall → apps → snapshots / Auto Backup.
Local AI on Contabo
RAM, disk floors, CPU Core vs Contabo GPU VPS, runtime patterns.
AI automation pipelines
n8n, video/image jobs, queues, same-box vs split-box.
Billing, backup, cancel
Pay, renew, cancel/refunds, pause AI before backup.
How clusters work
Use-case hubs (n8n, Local AI, WordPress) teach procurement without locking you to one brand. Provider clusters answer “what is it like to operate this vendor specifically?” They share quality standards, not a cloned template: commands and recovery advice follow each product’s documented controls.
Next providers
Kamatera is now live alongside ScalaHosting and prior clusters. Further provider clusters will follow fresh provider-specific research and positioning. Shared quality standards do not mean cloned commands, recovery assumptions or purchase claims.