This document answers the decision question — which provider, which tier — not how to install. The complete installation and first-connection walkthrough lives in Guides, which covers each step in order: where to click, what you should see, and what comes next. This page breaks down the route types, bandwidth definitions, billing models, device limits and refund terms you need to understand before buying, and closes with a pre-purchase checklist. Use the two together: settle which metrics matter here, then get the service running on the Guides page.
The judgments in this document come in two kinds. One is general buying method that applies across the industry — the cost relationship between dedicated lines and relays, for example, or where peak-hour pressure comes from. The other is VPN31's specific terms and tiers, drawn only from the pricing page and the Terms of Service. The two are kept separate, and any figure about this service is one you can verify.
The VPN31 facts used on this page: 100+ countries / 230+ routes, unlimited devices, 14-day no-questions refund, monthly plans from ¥9.9, data packs from ¥158, sign-up without an email address, and support for Alipay / WeChat / USDT. No figures outside this list appear here.
First, define your needs: the three user types care about different metrics
The most common mistake when buying cross-border network service is to look at price before needs. Price matters, of course, but the same price buys very different things from different providers: some spend it on dedicated-line bandwidth, some on node count, some on apps and support. Pin down your own usage first, then compare metrics — the judgment gets much steadier.
Below, the three common user types are separated out. One person can belong to two at once — office work by day, streaming by night — in which case set the primary metric by whichever takes more of your time and treat the other as secondary.
Users whose workday runs on overseas services
These users spend a fixed number of hours every day connected to overseas services: email, meetings, code repositories, internal systems. For them stability outranks peak speed — one dropped meeting costs far more than a slightly higher monthly fee. The metrics to check are route type (is there a dedicated or relay option), whether routes stay open at peak hours, and whether reconnection after a drop is clean. Price sensitivity is lowest here but the cost of churn is highest, so read the refund terms carefully: if a route turns out to perform badly in your city, whether you can get out within the promised number of days decides how expensive that trial is.
Families and multi-device users
One account has to run on phones, tablets, laptops and a TV box at the same time, so device count becomes a hard requirement. Two things need confirming: whether 'unlimited devices' really means no per-device charge, and how bandwidth is divided once several devices are online together. Sharing thins out the bandwidth each device gets; if someone is watching 4K while someone else is in a meeting, the demand on route quality is clearly higher than for a single user. Look for dedicated or relay options first, then price.
Users who mainly watch streaming and use AI tools
Usage here is bursty: two hours of streaming in the evening, the occasional AI tool during the day. What matters is whether the target site opens reliably and loads cleanly, not round-the-clock presence. For streaming, the location and actual exit of the landing route matter more than the bandwidth figure; for AI tools, the stability of a long-lived connection matters more than peak speed — a generation that dies halfway is worse than one that runs a little slower overall. Peak traffic is low for these users but concentrated in the same hours, so peak-hour route quality decides almost the entire experience.
Four common misjudgments
- Price only. A low price usually means shared bandwidth and less route redundancy; the difference is invisible off-peak and obvious at eight in the evening.
- Node count only. Node count is the easiest metric to manufacture: several entry points on one machine double the number while the routes you can actually use may not increase at all.
- Treating speed-test screenshots as conclusions. Results depend on time of day, your local network and where the test server sits; someone else's screenshot tells you little.
- Ignoring the refund terms. The terms decide whether you can walk away once you find a problem, and they cut the cost of a wrong guess more than any marketing line.
| User type | Primary metric | Secondary metric | Where you can compromise |
|---|---|---|---|
| Daily overseas office work | Route type and peak-hour stability | Reconnection, support response | Peak bandwidth |
| Family sharing | Device count, bandwidth per device | Platform coverage | Node count |
| Content consumption | Availability of target sites | Price, data allowance | Round-the-clock presence |
Find your own row first, then use the later chapters to see how each metric should be judged.
Route types: how IEPL dedicated lines, relays and direct routes differ in cost and experience
Route type is the item most worth your time, because it determines both cost and experience — and it is verifiable information. Whether a provider dares to list the type of every route is itself a signal.
Direct: lowest cost, most exposed to time of day
Direct means traffic leaves your local network for the international gateway and reaches the target server without passing through any relay the provider runs. It is the cheapest option, and the route comes from the public internet, so once the cross-border gateway congests at peak hours the effect is immediate: higher latency, more packet loss, slower video start. Fine for latency-tolerant work such as web browsing, file downloads and non-real-time sync.
Relay: trading a detour for stability
A relay works by placing nodes in-country or near the border: you connect to the relay first, and the relay reaches the landing server over an optimized path. Cost sits in the middle, and the gain is skipping a congested stretch of the public internet. Relay quality depends on the bandwidth and scheduling of the relay nodes, and different relay routes from the same provider can vary widely — so check whether relays are labelled as such rather than lumped together as 'high-speed routes'.
IEPL: routing around the congestion entirely
International Ethernet Private Line is, in essence, renting a dedicated or semi-dedicated channel from a carrier that avoids the congested points of the public internet. It is the most expensive option, usually priced by bandwidth and city pair, which is why dedicated routes generally cost more. The gain shows up at peak hours: when the public gateway starts dropping packets, latency and jitter on the dedicated channel barely move. For meetings, remote desktop and real-time collaboration — anything sensitive to jitter — a dedicated line is worth far more than the extra peak bandwidth.
| Route type | Path characteristics | Relative cost | Peak-hour behavior | Best for |
|---|---|---|---|---|
| Direct | Local network straight to the international gateway | Lowest | Clearly affected by public congestion | Web browsing, file downloads |
| Relay | To a relay node first, then an optimized path to the landing server | Moderate | Steadier than direct; depends on relay bandwidth | Everyday office work, video |
| IEPL dedicated line | Leased dedicated channel that avoids public congestion points | Highest | Latency and jitter stay essentially flat | Meetings, remote desktop, real-time collaboration |
How to mix them within one service
A mature subscription service usually offers all three route types rather than just one. The sensible pattern: put the tasks that need the most stability on dedicated lines, put bulk downloads and streaming on relays or direct routes, and let the expensive bandwidth serve only the cases that truly need it. VPN31 labels every route in its list by country and city with its type (IEPL dedicated, relay, direct), so you can pick against it directly; the full list is on the routes page.
Route type is information that can be written into a table and checked; adjectives like 'optimized', 'high-speed' and 'blazing' cannot. Prefer the former when buying: a provider willing to label every route's type is at least verifiable on this point.
One more point that is often missed: the same route behaves differently in different cities. The stretch between the landing server and your home broadband is what decides your personal experience. So when you read anyone's recommendation, check whether their city and ISP are close to yours before copying it.
Bandwidth and concurrency: where marketing definitions diverge from real experience
Bandwidth is the item most easily skewed by numbers. Nearly every provider quotes a big one, but few explain what it measures. Different definitions cannot be compared.
Three definitions that get confused
Total node bandwidth. The ceiling one landing server can reach, shared by every user connected to it. A big number does not mean you alone can use that much; it only says how wide that machine's exit is.
Per-user speed limit. The cap a provider sets on a single connection or account. This is the number that directly concerns you, and few providers state it clearly, because it exposes real cost fastest.
Concurrent connections. How many devices or connections one account can have online at once. It affects whether you can all use it together, not the speed of a single device. On this item VPN31 is unlimited devices.
Mix the three definitions together and the bandwidth figure loses all comparative meaning. When you see a number, ask which category it belongs to before deciding whether it is worth reading.
Peak hours are the real test
Pressure on cross-border routes concentrates between 8pm and 11pm: public gateway capacity is taken up at the same time, public routes start dropping packets, and nodes on shared bandwidth start queueing. What you get in that window is your everyday experience; a daytime speed test proves nothing. To judge whether a provider is solid, see whether it states that routes stay open at peak hours, and whether dedicated or relay options exist.
Jitter and packet loss matter more than peak speed
A stuttering video call is usually not a bandwidth shortage but too much jitter; a page that will not load is usually not slow speed but packet loss forcing retransmits. So 'is latency steady' is worth more attention than 'how high does the peak go'. Dedicated lines beat the alternatives on those two measures by a wider margin than on peak bandwidth, which is why people who work online all day are often willing to pay a little more for them.
If a provider's only selling point is one large bandwidth number, without saying whether it is total node bandwidth or a per-user cap, or which hours it was measured in, that number cannot be used for comparison. Treat it as a reference, not a conclusion.
How to test it yourself
The most reliable method is to test with the services you actually use, not with a speed-test site. Concretely: pick a weekday peak window, three days in a row, same device and same route, open the two or three services you use most, and note how fast they open and whether streams break. Three days is enough to see a pattern, and enough to finish the judgment inside the refund window. The test steps and common failure causes are covered in the verification section of Guides.
One more trick while testing: switch to a route of a different type and test again the same day. The gap between dedicated, relay and direct in the same time window is more telling than any written description, and it tells you whether you actually need to pay for a dedicated line.
Billing models: monthly subscriptions versus data packs
The billing model decides whether you pay for time or for allowance. Choosing the wrong model wastes more than choosing the wrong tier: a tier off by one costs a little more, but the wrong model can let a whole allowance expire at month end.
Subscriptions buy time, data packs buy allowance
A monthly subscription bills by time and gives you a data allowance each month that does not roll over. A data pack bills by allowance: you use what you bought, until it runs out, with no expiry date. The first suits people who are online every day; the second suits people whose usage is uneven, or who just want a backup route on hand.
The three monthly tiers and their rules
Monthly subscriptions come in three tiers: ¥9.9/month with 60GB, ¥18/month with 250GB, and ¥28/month with 500GB. Allowances reset monthly on your activation date, not on the calendar month, so your cycle starts the day you pay. When you upgrade mid-cycle, the price difference is converted into remaining days — you do not lose the rest of the current month's allowance, which matters if you only discover halfway through the month that you need more.
The three data packs and their rules
Data packs come in three tiers: ¥158 for 300GB, ¥358 for 1000GB and ¥658 for 3000GB. They last until used up and never expire. They do not follow a monthly cycle, which suits business trips, short projects, or keeping a fallback on a main account. The allowance does not disappear as time passes — that is the biggest difference from a subscription.
| Billing | Tier | Allowance | Validity rule | Suits |
|---|---|---|---|---|
| Monthly subscription | ¥9.9/month | 60GB | Resets monthly on activation date | Light use, trying it out |
| Monthly subscription | ¥18/month | 250GB | Resets monthly on activation date | Everyday office work plus streaming |
| Monthly subscription | ¥28/month | 500GB | Resets monthly on activation date | Family sharing, multiple devices |
| Data pack | ¥158 | 300GB | Lasts until used up, never expires | Backup, concentrated trip usage |
| Data pack | ¥358 | 1000GB | Lasts until used up, never expires | Seasonal heavy usage |
| Data pack | ¥658 | 3000GB | Lasts until used up, never expires | Long-term backup |
Estimate your monthly usage first
Before picking a tier, estimate roughly how much you use in a month. It does not need to be precise — adding up the three main categories is enough:
Monthly data estimate template (fill in your own numbers)
Video per-hour usage × hours per month = ______ GB
Download file size × files per month = ______ GB
Everyday daily estimate × 30 = ______ GB
--------------------------------------------
Total = ______ GBAdd 30% headroom on top of the total and that is your tier. The reason: usage has surprise spikes mid-month — system updates, cloud backups, video defaulting to a higher resolution — all push actual use past the estimate. Headroom is cheap; upgrading after running out is not.
When to choose a data pack
Three situations suit data packs: usage concentrated in certain months, such as seasonal trips or projects; already having a main account and wanting only a backup route against surprises; and not yet knowing whether you will stay long-term, so buying one allowance to try. Conversely, if you are online every day with steady usage, a subscription has a lower unit price and less to think about.
Monthly subscriptions and data packs follow different rules: the former resets monthly on the activation date, the latter lasts until used up and never expires. Pick the one that matches your usage pattern; there is no need to buy both just to cover every case.
Payment supports Alipay, WeChat and USDT — use whichever is convenient. All tiers and current prices are on the pricing page.
Device count and family sharing: what unlimited devices actually means
'Unlimited devices' deserves its own section, because it is both the most concrete promise and the most misunderstood one. It says one specific thing: the account is not charged by the number of devices online at once.
No per-device charge does not mean unlimited bandwidth
Unlimited devices means you can sign in on phones, tablets, laptops and a TV box without paying extra for the additional devices. But route bandwidth is a shared resource: the more devices online at once, the less each one gets. Three devices belonging to one person and three devices in a family put very different load on a route — the former mostly has one in use, the latter may have three streaming at once.
Platform coverage and installation differences
Five platforms are supported: Windows / macOS / iOS / Android / Linux. Installation is not identical across them: desktop systems usually require authorizing a network extension, macOS can conflict with iCloud Private Relay, and iOS needs you to allow the VPN configuration before it takes effect. These differences are covered step by step in Guides, and macOS permissions and chip compatibility have a separate write-up you can check against: permissions and compatibility tested on Mac.
Account security when sharing
Subscription details are account credentials: anyone who has them can consume your allowance. When sharing within a family, do not post the subscription link in public groups, and do not paste it into third-party tools for 'speed tests'. If a family member needs access, have them sign in on their own device rather than passing the link around. The more people use one account, the more this matters: if an outsider drains the allowance, all you see is data dropping faster than expected, with no obvious cause.
Sign-up requirements
VPN31 does not require an email address — a username and password are enough. That is especially handy for family sharing: no need to set up a mailbox for each member, and no getting stuck at step one because a verification email never arrives. For anyone who just wants the service running, this saves more time than you would expect.
Device count costs nothing extra, but route bandwidth is shared. Larger households should go straight for a higher allowance tier — what you save is not a device fee but queueing time.
Refunds and support: the terms to read before you pay
The refund terms are the first thing to read in this whole document, because they decide whether you can walk away once you find a problem. Prices can be compared and routes can be tested; only the terms must be read before paying.
The promise itself
VPN31 offers a 14-day no-questions refund. The full wording in the Terms of Service: within 14 days of the first payment you may request a full refund without giving a reason. The 'first payment' qualifier means it covers the first purchase, as a guard against a wrong guess, not an indefinite unconditional cancellation. The complete conditions and process are in the Terms of Service, which is worth reading through before you pay.
Why refund terms are a buying criterion
Cross-border route quality depends heavily on your city, your ISP and the hours you use — three things no review can test for you. The value of a refund policy is that it brings the cost of a wrong guess close to zero: you can run it in your own environment for two full weeks and walk away if it disappoints. More than any marketing copy, this lowers the risk of the decision. Use the same ruler on other providers: refund terms first, price second.
Support channels and response
If you hit a connection problem, use the ticket form in the user panel and include the route name, device platform, time of the error and the exact message you saw, plus a screenshot. The more complete the information, the faster the diagnosis. Tickets also leave a record: if the same problem recurs, the history helps you tell a route problem from a local network problem, and saves the person handling it several rounds of questions.
Payment methods and privacy
Payment supports Alipay, WeChat and USDT — use whatever suits you. On privacy, VPN31's position is no logging; the exact scope is set out in the Privacy Policy, including what is collected, what is not, and how long it is kept. With any provider, read both documents before paying: they say more about how a service treats its users than the homepage does.
Treat the refund window as a trial: buy one month, run your usual routes in your own city at your usual hours, and ask for a refund within 14 days if it disappoints. That beats reading ten recommendations.
Common pitfalls: overselling, inflated node counts and the warning signs of a service winding down
This section is about how to spot problems; it is not aimed at any particular provider. The criteria are general and can be applied to anyone, VPN31 included.
Overselling: selling more bandwidth than you own
Overselling is common industry practice; the question is the ratio. At a sane ratio users barely notice; at a high one, every route slows down together at peak hours while daytime stays normal. The test is concrete: measure the same route on several consecutive evenings. If speed drops consistently to a fraction of its daytime value and switching to another route restores it, that route's bandwidth is almost certainly over-allocated. If only one route is slow while others are fine, it is more likely a problem with that route — switch landing servers.
Inflated node counts: the easiest number to manufacture
Node counts get padded in several ways: counting a relay entry point as a node, counting several entry points on one landing machine as several nodes, or blurring 'countries covered' with 'countries with landing servers'. The test is simple: ask for a route list by country and city, and check whether each entry has an explicit route type. A total with no breakdown cannot be verified no matter how large it is. The routes page groups cities and route types by country, so you can check them one by one.
Warning signs of a service winding down
Outages usually cast signs beforehand: announcement updates slow noticeably, ticket responses stretch from hours to days, payment channels change frequently, plan prices drop oddly. Any one of these alone proves nothing; two or three together are worth heeding. The test is plain: look for traces of normal maintenance over the past three months — are announcements updated, have routes been added or removed, has the documentation kept pace.
Why low prices are not sustainable
Dedicated-line bandwidth is a fixed monthly cost that does not get cheaper no matter how cheaply it is resold. When a price sits clearly below route cost, only three things are possible: downgrading to cheaper routes, raising the oversell ratio, or funding existing users with new users' payments. All three eventually show up in the experience; only the timing differs. Do not chase the lowest price — chase a price that matches the route type.
Three general verification principles
- Ask for tables, not adjectives. Route type, city, allowance and refund days are facts that can be checked; adjectives are not.
- Ask for terms, not slogans. Refunds and privacy live in the terms, not the marketing copy; terms can be read word by word, slogans can only be believed or not.
- Start small. Verify on the shortest cycle first, then decide whether to stay. This step filters out most bad calls.
The pre-purchase checklist and decision flow
The previous sections compressed into a checklist you can follow in order. The order matters: scenario, then route, then allowance, then price. Price comes last not because it is unimportant, but because the first four decide what that price actually buys.
Six checks before you order
- Scenario. Decide whether you are office work, family sharing or content consumption — each has a different primary metric.
- Route. Does the route list label IEPL dedicated, relay and direct routes by city, and can you pick between them?
- Allowance. Estimate your monthly usage, add 30% headroom, then choose a subscription or a data pack.
- Devices. Confirm unlimited devices, and think through how many will be online at once and whether they will compete for bandwidth.
- Refund. Confirm the number of refund days and the conditions, and read the terms before paying.
- Payment. Confirm your usual payment method is supported, so it is not a surprise at checkout.
Verification steps during the trial
- Start on the shortest cycle; do not open with the longest tier.
- Test the two or three services you use most during peak hours (20:00–23:00), three days running.
- Switch to another route and retest, comparing dedicated, relay and direct.
- Install it on all your usual devices and confirm platform compatibility and permission settings.
- Log problems: route name, time, error symptoms — and open a ticket if needed.
- If it disappoints, request the refund within 14 days; do not leave it to the last day or two.
Decision path
If you remember one line: verify the route on the shortest cycle first, then choose the tier from what you learned. Doing it the other way round — buying a long plan and finding the problem afterwards — throws the refund terms away. Upgrading after verification converts the price difference into remaining days, so you do not lose out.
For worked examples, read the trade-off analysis by budget tier, or the ten questions beginners ask most. Both are written from the same set of facts and can be cross-checked against this page.
FAQ at a glance
These are the questions asked most often at the buying stage; each gives the conclusion first, then the reasoning.
Monthly subscription or data pack — which should I buy first?
Online every day with steady usage: choose a monthly subscription. Usage concentrated in certain months, or wanting only a backup route: choose a data pack. The rules differ: subscription allowances reset monthly on the activation date, while data packs last until used up and never expire.
What if I want to upgrade mid-cycle?
Monthly subscriptions support mid-cycle upgrades; the price difference is converted into remaining days, so the rest of the month's allowance is not lost. That makes buying the low tier first and upgrading once the routes check out the cheapest way in.
How does the 14-day no-questions refund work?
Within 14 days of the first payment you can request a full refund without giving a reason; the complete conditions and process are in the Terms of Service. Treat the refund window as a trial: run the routes in your own city at your usual hours before deciding.
How many devices can one account use?
Unlimited devices, with no extra charge per device. Note that route bandwidth is shared: the more devices online at once, the less each one gets.
Which platforms are supported?
Windows / macOS / iOS / Android / Linux — five platforms. Authorization and permission settings differ slightly between them; step-by-step instructions are in Guides.
Do I need an email address to sign up?
No. A username and password are enough, with no verification email to wait for. That is especially convenient for family sharing, since you do not need a separate mailbox for each member.
Are routes throttled at peak hours?
Routes stay open at peak hours. Actual speed depends on the route type you choose and your local network: dedicated lines hold up best at peak times, relays next, and direct routes are most exposed to public congestion. That is the main reason route type deserves the most weight.
How do I tell whether a route suits me?
Test in your own city, at your own hours, with the services you really use, over several days — then switch to a route of a different type and compare. If it disappoints, refund it within the window rather than putting up with it.
Which payment methods are supported?
Alipay, WeChat and USDT. All tiers and prices are on the pricing page.