MyCheap.Host is a budget hosting provider offering VPS, dedicated servers, proxies, GPU instances, CDN, and cloud storage, with server locations in France, Canada, Hong Kong SAR China, and India.
The platform is built around affordable pricing, anonymous payment options, and a no-activity-log privacy policy, targeting buyers who want capable infrastructure, on Linux or Windows, without the overhead of enterprise pricing or traditional billing requirements.
This MyCheap.Host review focuses on the VPS line, and whether that value proposition holds up under real-world testing is what it answers.
The scores below reflect hands-on testing of a France VPS covering the full Linux benchmark suite, the complete registration and ordering flow, and a direct ticket support interaction.
Performance scores are based on real test data rather than stated specifications, and ease of use scores reflect the experience of a first-time buyer going through the process from scratch. Scores are assigned using our HostAdvice rating methodology.
| Parameter | Score | Why This Score |
|---|---|---|
| Prices | 9.0/10 | Competitive at the budget tier; crypto avoids surcharges, but no money-back policy for provisioned services. |
| Features | 9.1/10 | NVMe standard, GPU available, Linux and Windows OS options, privacy-forward; no knowledge base. |
| Performance | 8.9/10 | Network exceptional at 4+ Gbps and zero failures under stress; CPU constrained, disk IOPS variance is high. |
| Ease of Use | 8.9/10 | Registration is fast and clean; pre-funding requirement is not communicated before the order error. |
| Support | 9.3/10 | Eleven-minute response on a technical question; single-sentence answer, no self-help options anywhere. |
| Overall | 9.0/10 | A stable, affordable VPS with a strong network, entry-level CPU limits, and a privacy edge. |
MyCheap offers six product types from a single account: VPS, Dedicated Servers, Proxies, GPU instances, CDN, and Cloud Storage.
The focus of this review is the VPS line, which runs from entry-level single-core instances through to 8 vCPU configurations, all backed by NVMe storage and 10 Gbps network links.
A few things to know before you top up:
Refund policy: Refunds are only issued if a paid service has not been delivered in full. If your VPS is provisioned and operational, you are not entitled to a refund. There is no money-back window for new customers evaluating the platform.
Payment methods:
Crypto is the most cost-effective option for international buyers. All fees are applied at the top-up stage, not at checkout.
Pre-funding model: You must add funds to your account balance before placing an order. There is no option to pay at the point of checkout. First-time buyers who skip the top-up step will hit an “Insufficient funds” error when they click Order Now.
Privacy: MyCheap accepts anonymous payment systems and does not save user activity logs. For buyers who want to minimize their personal data footprint, the crypto payment route paired with this no-log policy is a meaningful combination.
MyCheap’s performance section covers one VPS instance provisioned in France, their primary European location.
I ran the full Linux benchmark suite: sysbench for CPU and memory, fio for disk I/O, the Ookla speedtest CLI for network, and stress-ng for a sustained 3-minute load test across all three resource types.
The aim is not just to report raw numbers but to translate them into what they actually mean for real workloads. A benchmark figure on its own tells you nothing useful. The question worth answering is: what can you realistically run on this server without hitting limits?
Test instance:
| Parameter | Details |
|---|---|
| Instance name | HA-MyCheap-FR |
| Plan | 2 vCPU / 4 GB RAM / 20 GB NVMe |
| CPU | Intel Xeon Gold 6230 @ 2.10 GHz |
| OS | AlmaLinux 8.10 (Cerulean Leopard) |
| Kernel | 4.18.0-513.9.1.el8_9.x86_64 |
| Virtualization | KVM (confirmed from BIOS model) |
| Region | France |
| Swap | None configured |
One detail to flag before the results: no swap is configured on this instance.
With 3.8 GiB of usable RAM, any workload that exhausts memory will kill processes rather than fall back to disk. For production deployments, adding a swap file is worth doing before you go live.
I ran sysbench at two levels: single-thread to measure raw per-core throughput, and multi-thread across both vCPUs to see how they scale under load. The prime number limit was set to 20,000 throughout.
Results:
| Test | Events/sec | Avg Latency | 95th Pct Latency |
|---|---|---|---|
| Single-thread | 215.85 | 4.62ms | 5.18ms |
| Multi-thread (2 threads) | 436.58 | 4.57ms | 5.67ms |
| Thread fairness stddev | 49.50 | — | — |
The multi-thread result scales almost exactly 2x the single-thread figure (215 x 2 = 430 expected, 436 delivered). Both cores are pulling their weight, and neither is sitting idle.
The number that demands explanation is the single-thread result: 215 events/sec. The Intel Xeon Gold 6230 is a Cascade Lake processor that should deliver well over 1,000 events/sec in a single-threaded sysbench run on a lightly loaded host.
Getting 215 means this VPS is delivering roughly 15 to 20% of what the underlying CPU is capable of.
The most likely cause is CPU overcommitting on the host node. At MyCheap’s price point, allocating more vCPUs than the host has physical cores is a standard way to keep costs low. The consequence is that your vCPU competes for time on a physical core rather than owning one.
What this means for real workloads:
The thread fairness standard deviation of 49.50 across two threads points to some unevenness between the two vCPUs. On its own it is not alarming, but combined with the low throughput it reinforces the picture of a host under shared pressure.
I used sysbench with a 1K block size and a 10 GiB total transfer for both sequential write and sequential read.
Results:
| Test | Throughput |
|---|---|
| Sequential Write | 2,107.27 MiB/sec |
| Sequential Read | 2,976.20 MiB/sec |
Average latency on all memory operations registered at 0.00ms at the millisecond level, meaning memory access happens faster than the measurement can record.
These are reasonable numbers for a shared VPS. Not exceptional, but adequate. A Redis caching layer, PHP session storage, or a small in-memory database will have no trouble here. The memory subsystem can serve data faster than most applications at this scale will demand.
For a floor and ceiling reference: a heavily overloaded VPS would struggle to sustain even 1,000 MiB/s, while a dedicated server would push 30,000 MiB/s or more. At 2,107 MiB/s write and 2,976 MiB/s read, MyCheap’s France VPS sits in a reasonable middle range for its price tier.
I used fio for three scenarios: sequential write, sequential read, and random 4K mixed read/write under concurrent load.
Results:
| Test | Throughput | IOPS | IOPS Stddev | IOPS Min | IOPS Max |
|---|---|---|---|---|---|
| Sequential Write | 573 MiB/s (600 MB/s) | 572 | 237.23 | 111 | 1,009 |
| Sequential Read | 1,085 MiB/s (1,138 MB/s) | 1,084 | 242.18 | 664 | 1,568 |
| Random 4K Read | 19.0 MiB/s | 5,117 | 413.30 | 1,186 | 8,514 |
| Random 4K Write | 20.0 MiB/s | 5,124 | 420.23 | 1,098 | 8,468 |
The headline figures look strong, particularly the 1,085 MiB/s sequential read. For a plan at this price, those numbers are above average and suggest genuine NVMe storage rather than a slower SSD tier dressed up in marketing.
The column that tells the more honest story is the standard deviation. A stddev of 237 on sequential write against an average of 572 IOPS is extremely high, and the raw min and max confirm it. Write IOPS ranged from 111 at the low end to 1,009 at the high end within a single test run. That is a nearly 10x range.
In practical terms, this means disk performance is tied to what else is happening on the host node at any given moment. During a quiet period, writes are fast. During a busy period, they can drop to a fifth of the average.
For a web server or a low-traffic blog, you may never notice. For a database, a file processing pipeline, or anything that depends on I/O consistency, this variance is a real operational risk.
The random 4K IOPS at roughly 5,100 on both read and write is a decent result. MySQL or PostgreSQL under moderate query load should perform adequately. The same caveat applies: the minimum IOPS of around 1,100 seen during the test is the number to plan around, not the average of 5,100.
I ran the Ookla speedtest CLI twice against the same Paris test server. Both runs were needed to get a reliable picture.
Results:
| Run | Download | Upload | Idle Latency | Packet Loss |
|---|---|---|---|---|
| Run 1 | 4,195.91 Mbps | 4,614.57 Mbps | 0.55ms | 76.7% |
| Run 2 | 4,406.19 Mbps | 5,194.77 Mbps | 0.86ms | 0.0% |
Run 1 opened with a connection refused error from the test server before completing the transfer.
The 76.7% packet loss on that run follows directly from that error, not from MyCheap’s network.
Run 2 against the same server came back entirely clean, which is the result I treat as representative.
The raw bandwidth figures on both runs are exceptional. Over 4 Gbps download and 5 Gbps upload on a VPS at this price goes well beyond what most workloads will ever need. A busy web server, a game server, or a file hosting service would all sit comfortably within what this link can carry.
Idle latency of under 1ms on both runs confirms close proximity to the Paris test infrastructure. For serving users across Western Europe, that is a strong baseline.
The stress test forces each resource type to sustain maximum load for 3 minutes, the most useful check for catching instability that short benchmark bursts do not reveal.
Results:
| Stressor | Bogo Ops | Bogo Ops/s | Workers | Result |
|---|---|---|---|---|
| CPU (180s) | 161,816 | 898.95 | 2/2 | Passed, 0 failed |
| Memory (180s) | 9,188,567 | 51,023.14 | 2/2 | Passed, 0 failed |
| Disk I/O (180s) | 2,167,524 | 12,010.26 | 2/2 | Passed, 0 failed |
Zero failures across all three stressors. Every worker dispatched completed its full run and came back as passed, with no untrusted metrics flags.
This is the result that matters most for production reliability. A server can look fine in short tests and then fall apart under sustained pressure: throttling, dropping workers, or producing corrupted results. MyCheap’s France VPS did none of that. It held steady across CPU, memory, and disk for three full minutes on each stressor.
The CPU stress bogo ops/s of 898.95 lines up with what the low single-thread sysbench result predicted. The server is not fast. But it does not degrade under load, which is a more important property for production hosting than raw throughput.
MyCheap’s France VPS delivers the performance profile you would expect at this price: stable, adequate for light workloads, and clearly bounded at the top end.
The CPU is constrained by host overcommitting, disk speeds are fast on average but inconsistent under varying host load, and network bandwidth is a genuine strength.
One thing to be clear about: Every number in this section describes the 2 vCPU / 4 GB entry-level plan I tested, not MyCheap’s larger configurations. If your workload is a small website, a development environment, a proxy, or a lightweight application, this plan will get the job done. If you need sustained CPU throughput or predictable disk I/O for a production database or a busy application server, this configuration lacks the power for that by design. Start at the 4 vCPU tier or above and benchmark that plan directly before committing.
MyCheap positions itself as a budget-first hosting provider where speed matters more than hand-holding.
To find out whether that extends to the actual buying experience, I went through everything from scratch: landing on the homepage for the first time, creating an account, working through the dashboard, and attempting to order a server.
The process has some genuine strengths and a few friction points that prospective buyers should know about before they commit.
The homepage lands cleanly. MyCheap leads with a dark purple, neon-accented design that wastes no space: the navigation across the top lists AI tools, a Checker, Promotions, a Game Server section, Partners, and a Blog, which tells you something about who the platform is built for.
A featured VPS plan sits right in the hero section with specs and pricing already visible, so you can get a sense of what you are buying before you click anything.
Clicking Login opens a modal overlay rather than redirecting you to a dedicated registration page. The modal has two tabs at the top, Sign In and Sign Up, and keeps you on the homepage throughout.
It’s a compact approach, though some users may find a full-page form more reassuring for submitting account credentials.
The Sign Up form asks for four fields:
Below the fields sit two checkboxes:
That pre-checked marketing checkbox is the kind of thing worth catching before you submit the form.
You can also register via Google or Telegram. The Google option is standard.
Telegram as a social login alternative is less common and signals something about MyCheap’s user base. Most Western-facing providers pair Google with GitHub or Apple.
Telegram suggests a meaningful share of MyCheap’s customers come from markets where it functions as a primary communication layer.
After submitting, MyCheap sends a 6-digit verification code to your email. The code input screen is clean and functional.
The one friction point in the registration flow is what happens after you enter the code. Instead of logging you in automatically, MyCheap redirects you back to the Sign In modal and asks you to log in again.
Your email is pre-filled, which saves some typing, but the extra step is unnecessary. A verified account should land in the dashboard without a second sign-in.
The full process from the homepage to an active session took me around three minutes. No billing details are required before sign-up, no plan selection is forced during registration, and the form is minimal enough that most users will move through it quickly.
After signing in, MyCheap brings you to a single interface that acts as both your account dashboard and the place where you purchase services. There is no separate client portal or management panel to switch between.
The layout is organized without being crowded. A vertical icon sidebar on the left gives you access to all account sections:
The top bar shows a Services dropdown, a currency selector, your account balance, a language toggle, and Logout. The currency selector supports USD, EUR, and RUB. The inclusion of RUB is a meaningful detail.
It’s not a negative on its own, but it places MyCheap’s origins clearly. For buyers paying in USD or EUR, you are accommodated, but the platform was not designed with a Western audience as its primary target.
The main body of the dashboard has two areas. On the left, your account balance with TOP UP and HISTORY buttons and a Finance chart that is empty for new accounts.
On the right, a grid of service tiles, each with a direct action button.
Moving from the dashboard to an order takes one click from any of these tiles.
The first thing visible on login after the balance widget is a review discount banner. MyCheap offers up to 10% off your first purchase by leaving reviews across four platforms at 2.5% per platform.
For someone who wants to get a server ordered quickly, the dashboard removes unnecessary steps. The combined client area and purchase flow works better in practice than it sounds in description.
Clicking the VPS tile takes you to a Buy VPS page built around a world map showing server location dots.
A Configuration panel on the right side lets you filter by country, billing period, and promo code before the plan list loads below.
The country dropdown shows all listed regions alongside their availability, including the ones you cannot order:
| Country | Status |
|---|---|
| Canada | Available |
| France | Available |
| Hong Kong SAR China | Available |
| India | Available |
| Germany | Sold Out |
| Kazakhstan | Sold Out |
| Netherlands | Sold Out |
I appreciate that MyCheap shows unavailable regions rather than hiding them. That said, three of seven listed locations being sold out is a sign that capacity availability is something to verify before committing to a specific region.
After selecting France and keeping the billing period at one month, the plan list loaded below the map.
All plans carry the same generic label: vCPU, with no tier names or plan branding, just a spec row and a price.
Comparing on raw specs is easy. Referencing a specific plan later when managing multiple servers is harder without any naming to anchor to.
Each plan has an OS dropdown built directly into its row, which keeps configuration in one place rather than spreading it across separate steps.
The selection is wider than it first appears. The visible portion of the dropdown lists Linux distributions, including AlmaLinux 8, 9, and 10, and CentOS 7 through 10 Stream.
Windows Server is also available in various versions further down the list. If you are ordering a Windows VPS, scroll past the Linux entries or type Windows into the search field at the top of the dropdown to find it.
When I clicked ORDER NOW, the portal returned an “Insufficient funds” error.
MyCheap runs on a pre-funded balance model. You cannot pay at checkout. The balance in your account must cover the order cost before the system processes it. This is standard for budget VPS providers, but it is not communicated anywhere in the order flow. A first-time buyer who has not topped up will hit this error cold.
To add funds, you return to the dashboard and click TOP UP on the balance widget. The top-up modal offers three payment methods:
| Payment Method | Fee | Notes |
|---|---|---|
| YOOKASSA | +5% | Primarily for buyers in Russia and the CIS region |
| International Bank Cards | +3% | Available worldwide |
| Crypto | Blockchain network fees only | No additional charge from MyCheap |
For most international buyers, bank card or crypto are the practical routes. The surcharges on card payments are worth factoring into your total cost if you plan to top up in smaller amounts.
The Configuration panel also shows a promo code field with “START20” as a placeholder hint. Whether this code is active and applicable to new orders is worth confirming with MyCheap directly before checkout.
After adding funds, the order went through without any further issues.
MyCheap’s interface is built for speed rather than ceremony, and that shows up in both the good and the frustrating parts of the experience. Registration is fast, the dashboard is direct, and the path from account creation to VPS order is short once you know the one rule: fund your balance first.
The friction points are real and worth naming:
For price-sensitive buyers who want a VPS in France, Canada, Hong Kong, or India and are comfortable with the pre-funded billing model, MyCheap delivers what it promises without unnecessary complexity. For buyers who expect a more guided onboarding experience or need reliable access to a specific European region, the gaps here will likely show.
MyCheap’s ease of use is functional, lean, and occasionally rough around the edges. The registration and ordering flows move quickly, but the post-verification redirect, the undisclosed pre-funding requirement, and the sold-out regions create friction that a smoother setup would eliminate. Experienced VPS buyers will adapt quickly. Those newer to self-managed hosting may run into more unexpected stops than they would with a more polished provider.
MyCheap’s support offering is minimal compared to most providers at this tier. Before testing it directly, I mapped out what is actually available. The two channels on offer are:
There is no live chat in the portal, no phone line, and no knowledge base or documentation library anywhere on the site.
For a budget provider, a slim support lineup is not surprising, but it does mean ticket response time carries more weight. When a ticket is your only in-portal option, how fast and how well the team answers matters more than it would if you had alternatives to fall back on.
From the dashboard sidebar, clicking the Support icon takes you to the support tickets page. The page is empty by default, and the only available action is a + NEW TICKET button in the top right corner.
Clicking it opens a modal with a clean, category-based ticket type selector:
You pick the category that matches your issue, optionally link it to an active order, fill in a subject (255 character limit) and a message (3,000-character limit), and hit Send.
No department selection, no subcategory, no file attachment field visible. The form is simpler than most portal-based systems, which is not a complaint at this price point. It is quick to fill out, and the category selection means the right team should receive it without triage delays.
I selected VPS as the ticket type, left the order field blank, and submitted a two-part technical question at 13:52 on 27 July 2026: what hypervisor technology does MyCheap use, and are CPU and RAM resources fully dedicated per instance or overcommitted across the host node.
After submitting, the ticket appeared in the Support page as ticket number 95, status Open, timestamped 13:52.
The reply arrived at 14:03, eleven minutes after submission. The agent is identified only as “Support” with no name or title attached. The full reply was a single sentence:
“Hello. We use KVM; we do not oversell, and all resources are fully dedicated.”
Eleven minutes for a two-part technical question is a fast result. Most budget providers take hours on a first reply. MyCheap beat that by a wide margin.
The content of the answer is where the review gets more complicated. The KVM confirmation is accurate and consistent with what the benchmark data already showed independently: the BIOS model on the provisioned France instance displayed the i440FX + PIIX chipset, which is characteristic of KVM and QEMU virtualization. On that point, the agent got it right.
The second claim, that resources are not oversold and are fully dedicated per instance, is harder to square with the performance numbers.
A Xeon Gold 6230 running a single-threaded sysbench test on a host with properly dedicated vCPUs should return well over 1,000 events/sec. The France VPS returned 215 events/sec, roughly 15 to 20% of that figure. That gap does not appear on a host where CPU time is properly allocated one-to-one.
There are two reasonable readings of this. One is that the agent gave an accurate statement about MyCheap’s stated platform policy, and the CPU result reflects something situational such as a heavily loaded host node at the time of testing.
The other is that this was a surface-level answer given without checking the actual host configuration. Both are possible. Readers who need reliable CPU throughput should treat the benchmark result as the primary signal and run their own tests rather than relying on the agent’s assurance alone.
One additional gap: the response addressed both questions in a single sentence but with no technical depth on either. A more thorough answer might have named the overcommit ratio, described how allocation works in practice, or asked a clarifying question about the workload. What came back was a one-liner. For basic account questions that is fine. For infrastructure planning decisions, it leaves work for the buyer to do.
There is no knowledge base or documentation library visible anywhere on MyCheap’s site or within the portal. Self-help is not an option here.
Every question becomes either a ticket or a Telegram message. For an experienced Linux user who rarely needs help, this is unlikely to matter. For someone managing their first VPS, it means no safety net between running into a problem and waiting for a human reply.
MyCheap’s support is fast and minimal. The ticket system is easy to access, the eleven-minute response time on a technical question is above average for this price tier, and the KVM confirmation was accurate.
The resource allocation claim conflicts with the CPU benchmark data, which means the support team may not always have full visibility into host-level infrastructure detail when they answer. For account and billing questions, the channel works.
For production infrastructure decisions, the benchmark results are the more reliable reference.
Yes, with a clear set of conditions attached. MyCheap delivers on the things it does well: the France VPS held up under sustained stress with zero failures, the network hit over 4 Gbps on a clean run, and the ticket support came back in eleven minutes with an accurate answer on hypervisor type. At the price being charged, that combination is real value.
The caveats matter as much as the strengths, and they apply to the specific server I tested, not the full lineup. The 2 vCPU entry-level plan returned 215 events/sec on a processor that should deliver over 1,000 events/sec under properly dedicated allocation. The support agent stated resources are fully dedicated, but the performance data on this instance does not align with that claim. Disk IOPS variance was also wide enough that a production database would need retesting before any serious deployment.
The plan I tested is the right pick for buyers who need a lightweight VPS for a website, a proxy, a game server, a development box, or a privacy-sensitive workload where anonymous payments and a no-log policy matter. The network link puts it ahead of many providers in the same price range.
That same entry-level configuration lacks the power for large-scale projects by design, so CPU-intensive workloads and production databases that depend on consistent I/O belong on MyCheap’s 4, 6, or 8 vCPU tiers, which I did not benchmark in this review. Separately, anyone who needs a refund window on a provisioned service should know there is none. Three of the seven listed regions were also sold out at the time of review, so confirm your target location is available before signing up.
MyCheap.Host is a competitive budget hosting provider offering VPS, dedicated servers, proxies, GPU instances, CDN, and cloud storage, with fast network speeds, KVM virtualization, and a privacy-focused approach. The 2 vCPU entry-level plan tested handles lightweight workloads well, though its CPU performance came in below the hardware’s potential. Resource-intensive projects belong on the higher-tier plans, which were not benchmarked in this review.
MyCheap.Host has server locations in France, Canada, Hong Kong SAR China, and India. Germany, Kazakhstan, and the Netherlands were listed as available regions but marked as sold out during testing.
No. MyCheap.Host only issues refunds if a paid service has not been delivered in full. Once a service is provisioned and running, there is no money-back window. If you want to test before paying, the free one-day proxy trial and one-week CDN trial are the only no-cost entry points, as VPS plans carry no trial.
MyCheap.Host uses KVM (Kernel-based Virtual Machine) for VPS instances. This was confirmed by their support team and independently verified through the BIOS model on the provisioned France server during benchmarking.
MyCheap.Host does not save user activity logs and accepts anonymous payment methods including cryptocurrency. For buyers who want to minimize their personal data trail, the combination of a no-log policy and crypto payment support makes MyCheap a practical option.
Answer a few simple questions and find the perfect solution for you!
Start Hosting Search
