Quick answer
Should I choose rotating or sticky African proxies?
Choose rotating African proxies for independent authorised observations; use sticky mode when a short multi-step test needs continuity. As of 15 October 2026, AfricaProxies residential traffic starts at $0.50/GB and sticky sessions run up to 30 minutes, with earlier changes possible. Connection reuse and browser cookies still need separate control in Nigeria, Kenya or South Africa comparisons.
The supported session rules
| Rotating mode | Rotating mode: each new connection gets an IP from the pool, usually a different one. To get a new IP on every request, send one request per connection (turn off keep-alive). Sticky mode keeps one IP for the session time you set. |
|---|---|
| Sticky ceiling | Sticky sessions keep the same IP for the time you set (ttl, in minutes), up to 30 minutes. The IP can change earlier, for example if the residential device goes offline. |
| Targeting | Country and city targeting on every plan; in the United States also state and ZIP code; ASN targeting at no extra cost. City availability varies. |
| Billing | Usage counts all data sent and received through the proxy (upload and download, headers included). 1 GB = 1,000,000,000 bytes. Connection attempts that fail before reaching the target are not counted. |
How do I choose and validate a proxy session mode?
Define continuity
Choose rotation for independent observations or sticky mode for a short sequence in your own application. Keep the proxy setting separate from cookies, login state and delivery settings.
Apply the documented mode
For independent requests, account for connection reuse and the keep-alive rule below. For sticky mode, set ttl in minutes up to 30; the exit can change earlier.
Record the actual exit
Use one observed address for the location check. Save any session interruption with the result, and keep an unavailable city distinct from a country-level fallback.
Which mode fits the measurement?
| Business observation | Starting choice | What to preserve |
|---|---|---|
| Independent public product samples with permission | Rotating connections | Exact variant, seller, delivery setting, currency and sample time |
| Short multi-step QA of your own application | Sticky session | Session ID, app cookies, test fixture and interruption state |
| Compare a single exit in two geolocation datasets | Sticky, or look up one recorded address | The same address and both dataset names |
| Nairobi landing page followed by an application step | Sticky if the brief needs continuity | City verification, browser language and the approved destination |
Why is a request different from a connection?
An application request is a page or API operation. A transport connection can carry more than one request, and HTTPS commonly uses a proxy CONNECT tunnel. A client that reuses its connection may therefore keep an exit across several operations. Count observed connections and addresses rather than assuming that a loop of page requests proves independent routing.
If independent observations are needed, configure the client according to the rotating rule above and inspect the resulting exits. A small available pool may return an address seen before. Repeated addresses do not, by themselves, establish that a different country was selected or that a particular ISP is unavailable.
When does sticky mode help an African application test?
Consider a QA flow in an application you own: open a catalogue, select a delivery region and inspect a cart fixture without placing a real order. A sticky proxy session reduces one changing input while the test inspects its expected language and currency. Keep that session separate from a browser cookie jar, application login and saved delivery address.
Run the same documented flow for South Africa and Kenya using separate settings and records. A sticky address does not transform a rand fixture into a shilling fixture or make two different delivery regions comparable. If the residential device disconnects, stop the continuity test and mark it interrupted instead of reporting an application defect.
How do I validate a city without mixing exits?
For a city-sensitive Lagos or Nairobi study, obtain one observed address from a controlled endpoint and check that address in the location sources you selected. Looking up two independently detected connections in rotating mode can accidentally compare two different exits. Keep the city requested, city returned, source names and time with the observation.
A country-level fallback can still be useful for a country-level question, but label its scope explicitly. If the brief requires Nairobi, the correct result when suitable exits are unavailable is an unavailable city sample. It is not a negative conclusion about local campaign delivery or an assurance that the fallback represents Nairobi.
How should session changes affect a retail study?
A price study should keep its matching rules consistent across samples: model, variant, seller, quantity, currency, delivery location and promotion conditions. Decide in advance whether each observation is independent or part of a short sequence. Record a session interruption beside the data rather than quietly repeating the operation until a preferred result appears.
A target refusal or rate limit ends the permitted request until its access conditions are reviewed. Session selection supplies a network method; it cannot authorise collection, remove a deliberate target restriction or validate the reuse of personal information. The retail guide describes the separate record-matching work.
Session questions
Does sticky mode guarantee an IP for thirty minutes?
Sticky sessions keep the same IP for the time you set (ttl, in minutes), up to 30 minutes. The IP can change earlier, for example if the residential device goes offline. Treat the ceiling as a maximum request for continuity and design your own application test so an early change is identifiable. Preserve test state and record the interruption before restarting under the conditions your brief allows.
Do I pay a separate fee to select a session?
Session mode is part of residential traffic access. The selected package is billed by its measured proxy traffic rather than by how many browser tabs you open. Start with the current pricing and counting policy, then estimate your Nigerian or Ghanaian job from a small permitted sample.