Proxies for Dolphin Anty

Setup takes a couple of minutes and needs no server configuration on your side:

1. Copy the access string from your IPrazon dashboard. The main format is IP:PORT, or IP:PORT:LOGIN:PASS if you need it.
2. Open a profile in Dolphin Anty or create a new one.
3. Choose the proxy type: SOCKS5 or HTTP, both are in the package.
4. Paste the data into the fields: host, port, login, password.
5. Click Check: Dolphin shows the country and connection status.
6. Save the profile: it now always goes out through the assigned IP.

For teams it is convenient to distribute profiles across proxy pool addresses and rotate IPs on schedule or on demand, which lowers the load on each address.

The starting point is simple: one profile should go out consistently through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and through one proxy and from one proxy address, otherwise the separation loses its point. So the lower bound is as many addresses as there are profiles running at the same time.

The gap between those two numbers is usually wide. Out of a hundred created profiles only a handful are active at any moment, so the proxy pool is sized for peak concurrency. The practical way to count it: look at how many windows are genuinely open in your busiest hour and leave a small margin above that.

The second factor is session length. If a profile runs for hours, the address is occupied all that time. If the task is short, one proxy address serves several profiles in sequence and the total number you need falls.

A refusal on connection almost always comes down to four causes, and checking them in this order saves time. First: the wrong proxy type is selected, HTTP where SOCKS5 is needed or the other way round. Second: the port does not match the one issued.

The third cause sits in the format. If the list is issued as IP:PORT with no login, access works by binding, and requests only go through from the address you registered. Changing machine, network or outbound interface returns a refusal on a perfectly working list.

Fourth: a typo in the login or password, most often a trailing space picked up when copying the string. Check separately that Dolphin Anty actually routes traffic through the proxy: a connection is confirmed by the source address on an IP test service.

Address switching is configured in two places and mixing them up causes trouble. The first is the proxy pool side: the address changes on its own by schedule or per request, and Dolphin Anty needs to do nothing. The second is the software side: you feed different strings from the list to different tasks yourself.

The first option suits uniform work where no state is held between requests. The list is connected once and new proxy addresses arrive automatically, with no intervention needed.

The second option is what you need when a profile holds an authorization and must go out from one address for the whole session. The address is then pinned explicitly and you decide when to change it: after a cycle finishes, after a number of processed items, or on a timer.

Running both modes on one task is usually harmful. If the proxy pool rotates on its own while the software cycles through the list in parallel, behaviour becomes unpredictable and some requests leave from addresses you did not intend.

The base rule: load has to fall evenly across the proxy pool, otherwise some addresses burn out ahead of the rest. So it is better to assign profiles to addresses deliberately than to let the software grab whichever comes first.

When tasks differ in weight, count by intensity. One heavy process can produce more requests than five light ones, and it makes sense to give it dedicated addresses so it does not crowd the others.

Keeping a small reserve of addresses that are not permanently occupied pays off. It comes in useful when part of the proxy pool starts meeting checks: you move the affected tasks onto fresh addresses without stopping everything.

And finally, record which task ran from which address. Without that, when refusals appear there is no way to tell whether the cause is a specific address, the task settings or the platform itself.

The proxy pool runs on our own hardware: stable fast addresses, low latency, uptime without drops. For multi-profile work this is critical, a profile must not lose connection mid-scenario.

Rotation delivers new IPs on schedule or on demand, and traffic is unlimited on all packages: profile workload is not capped. The pool holds 12,000 addresses across 200+ countries.

SOCKS5 and HTTP are in every package, with login/password or IP authorization. That is enough to assign each profile its own address and keep profiles independent.