The short answer
- 1 NodeMaven Best clean-IP latency Visit site
- 2 Decodo (ex Smartproxy) Best city-level targeting Visit site
- 3 Novada Best mobile carrier spread Visit site
17 best proxies for sneakers, compared
| # | Provider | Fit | From | Readers | Protocols | Links |
|---|---|---|---|---|---|---|
| 1 | 4.8/5 | $2.20 / GB | 88% | SOCKS5, HTTP, HTTPS | ||
|
Residential, mobile and ISP proxies with 95% clean IPs, residential fraud score under 5/100, unlimited concurrent sessions and country/city/ISP targeting. Good for this job
Watch out
|
||||||
| 2 | 4.7/5 | $0.08 / GB | 100% | HTTP(S), SOCKS5 | ||
|
115M+ residential IPs, city/ZIP/ASN targeting, sticky sessions up to 30 min, 99.92% residential success and sub-0.5s response. Good for this job
Watch out
|
||||||
| 3 | 4.6/5 | $0.50 / GB | 88% | HTTP(S) | ||
|
100M+ residential IPs with city-level targeting, 2M+ mobile IPs across 120+ carriers, unlimited sessions and sub-0.5s response time. Good for this job
Watch out
|
||||||
| 4 | 4.5/5 | $0.02 / month | 100% | — | ||
|
80M+ rotating residential IPs with country, state, city, ZIP and ASN targeting; 500–3,000 thread limits and 99.7% success. Good for this job
Watch out
|
||||||
| 5 | 4.4/5 | $0.35 / GB | 90% | HTTP, HTTPS, SOCKS5, HTTP/2 | ||
|
30M+ residential and 4M+ mobile IPs, country/region/city/ZIP targeting, unlimited sessions for approved use and rollover bandwidth. Good for this job
Watch out
|
||||||
| 6 | 4.3/5 | $2 / GB | 80% | HTTPS, SOCKS5 | ||
|
Residential and mobile IPs on one plan, unlimited concurrent requests, sticky sessions up to 7 days, but only 38+ countries and ~0.6s response. Good for this job
Watch out
|
||||||
| 7 | 4.2/5 | $0.06 / GB | 100% | HTTP, HTTPS, SOCKS5 | ||
|
86M+ residential IPs across 200+ countries with country/state/city targeting and a $69.68/day unlimited residential option. Good for this job
Watch out
|
||||||
| 8 | 4.1/5 | $0.90 / GB | 86% | HTTP, HTTPS, SOCKS5 | ||
|
100M+ residential IPs, country/state/city targeting and unlimited concurrent sessions; third-party tests show 0.52–1.00s responses and no mobile line. Good for this job
Watch out
|
||||||
| 9 | 4.0/5 | $0.30 / GB | 96% | HTTP, HTTPS, SOCKS5 | ||
|
47M+ rotating residential IPs, ~0.35s response time and $0.30/GB, though mobile covers only 20+ countries and city targeting is not stated. Good for this job
Watch out
|
||||||
| 10 | 3.9/5 | $5 / proxy | 85% | — | ||
|
Per-proxy billing at $5 residential or $15 mobile per month, 7M+ IPs and own infrastructure, but only 150+ locations and no city targeting stated. Good for this job
Watch out
|
||||||
| 11 | 3.8/5 | $0.40 / GB | 100% | — | ||
|
175M+ residential IPs and 195 locations plus dedicated ISP options, but pricing starts at $6/GB and city-level targeting is not stated. Good for this job
Watch out
|
||||||
| 12 | 3.7/5 | $2 / GB | 88% | SOCKS4, SOCKS5, HTTP, HTTPS | ||
|
20M+ residential IPs with country/city/ZIP/ASN targeting, unlimited concurrent connections and sticky IPs up to 72 hours; mobile is Poland/Ukraine only. Good for this job
Watch out
|
||||||
| 13 | 3.6/5 | $0.63 / GB | 86% | SOCKS5, HTTP, HTTPS | ||
|
Rotating residential, mobile and datacenter from $1/GB, country/city/ASN targeting and 95% clean IPs; founded 2024 and concurrency is not stated. Good for this job
Watch out
|
||||||
| 14 | 3.4/5 | — | 81% | SOCKS5, HTTP, HTTPS | ||
|
80M+ residential IPs with country/city/ISP/ASN targeting, but no free trial or refunds and no concurrency figure is published. Good for this job
Watch out
|
||||||
| 15 | 3.3/5 | ₽490 / day | 83% | SOCKS5, HTTP | ||
|
Mobile proxies with unlimited traffic, 53 countries and up to 250 simultaneous connections per proxy, but 3.24s response time is slow for drops. Good for this job
Watch out
|
||||||
| 16 | 3.1/5 | $2 | 82% | HTTP/S | ||
|
No concurrency limit, unlimited threads and 60-minute sticky sessions, but geography is vague, state IPs are US-only and static residential is unavailable. Good for this job
Watch out
|
||||||
| 17 | 2.2/5 | $1 | 82% | — | ||
|
Shared proxies put multiple users on the same IP, with no city targeting and a $10 minimum; weak for clean retail drops. Good for this job
Watch out
|
||||||
Ordered by how well each provider fits this job, from the figures the providers publish, our review scores and reader votes — see our methodology.
What a sneaker bot actually needs from a proxy
A bot is a program making HTTP requests to a retailer's product page, add-to-cart endpoint and checkout endpoint. Nothing more exotic than that. What sits between those requests and the stock is a bot-management layer — Queue-it, Akamai, DataDome, PerimeterX, Cloudflare and similar — and that layer scores the request before it ever reaches a warehouse.
The score is built from several things at once: the reputation of the IP address, the autonomous system it belongs to, the TLS fingerprint of the client, the header order, the cookie history, whether the session behaves like a browser, and how fast the requests arrive. A proxy only fixes the first item on that list. It gives each task a different egress address so that one block does not take out thirty tasks at once, and so the retailer cannot link thirty checkouts to one household.
So the requirements are narrow and specific:
- A distinct egress IP per task. Not per bot — per task. Six tasks leaving from one address is one task with six attempts, as far as the retailer is concerned.
- An address that is not already burned. Retail pools are shared. The address you are handed may have been used on the same site this morning by somebody else, and their behaviour is now part of its history.
- Room to open connections at once. A drop is a burst. If the plan caps you at ten concurrent connections, tasks eleven onwards wait, and by then the shoe is gone.
- Stability for the length of the session. A rotating address that changes between add-to-cart and checkout drops the cart. Cookie and session state are tied to the address that created them.
- Geography that matches the drop. Regional releases, region-locked shipping and local stock mean the address's city matters, not just its country.
What a proxy cannot do is fix a bad client. If the bot's TLS fingerprint says Python and its headers are in the wrong order, a clean residential address buys you one request before the challenge page appears.
Residential, ISP or datacenter for a drop
Three address types, three different jobs. The mistake is treating them as a quality ladder rather than as tools.
Residential
Addresses assigned by consumer ISPs to real homes. They carry the best reputation because they look like what they are — somebody's broadband line. The catch is that they are pooled and shared: thousands of users rotate through the same addresses, so reputation is collective and unpredictable. Rotating residential is well suited to monitoring, stock pinging, account creation and anything where you need a large spread of addresses briefly. It is a poor fit for the checkout itself, because the address must stay put for the length of the session and the pool may not let you hold one that long.
ISP (static residential)
Residential addresses hosted on a commercial basis, sold as dedicated or semi-dedicated and held for as long as you pay. You get residential-grade reputation with the stability of a datacenter contract. For drops, this is the usual choice: one address, one task, held from queue entry to order confirmation. It costs more per address than rotating residential, and availability in a specific city is narrower.
Datacenter
Cheap, fast, plentiful, and immediately recognisable. Retail sites keep aggressive lists of hosting ASNs, and many will challenge or block a datacenter address on sight regardless of its history. That makes them a poor choice for the checkout path. They remain useful for the work around the drop — pinging a stock endpoint that has no bot protection, checking whether a page has changed, testing your own configuration — where being blocked costs you nothing.
The awkward part: no address type is clean forever. A dedicated ISP address that works for six months can be flagged in a week if you hammer one site with it. Buy for the drop, retire addresses that start getting challenged, and do not assume an expensive tier is a permanent pass.
If you want to see how a cheap or free address behaves before spending anything, take one from the free proxy list and watch how quickly a retailer challenges it. That is the floor the paid tiers are being compared against.
How many addresses per task
The safe rule is one address per task, per drop. Not one per account, not one per shoe — one per task. If you are running ten tasks, buy ten addresses. Anything denser means two tasks share a reputation and a rate limit, and when one trips a challenge the other inherits it.
That rule has a cost, so people bend it, and the bending has a shape:
- High-value drops, low stock: one to one. There is no reason to save money here; the margin on a single pair covers the difference.
- Mid-value drops, larger stock: one address to two tasks is common. One to three works if the tasks are spread across minutes rather than seconds, because the retailer's rate limiter is counting requests per address per window, not per day.
- Monitoring and restock pinging: density matters far less. A handful of addresses rotating across many checks is fine, because a challenge costs you a poll, not a pair.
Concurrency is the other half of the sum. A pool of fifty addresses on a plan that allows ten simultaneous connections behaves like ten addresses. Read the concurrency limit before the pool size — it is the number that sets your ceiling.
How to load proxies into a bot
Most bots have a Proxies or Network tab with a bulk import field, and nearly all of them want one of two formats. Get this right once and keep a template.
Common accepted shapes:
host:port:username:password— the most widely supported, colon-delimited.username:password@host:port— the URL form, used when the bot builds a proxy URL internally.scheme://username:password@host:port— explicit scheme, needed when the pool mixes HTTP and SOCKS5.
Practical steps:
- Export the list from the provider as plain text and open it in a text editor, not a spreadsheet — spreadsheet software will mangle long credentials.
- Strip whitespace and blank lines, sort the list so duplicates sit next to each other, and remove them. Duplicates are the quietest cause of a whole task group failing together.
- Assign each address to exactly one task, in the bot's task editor, before the drop rather than during it.
- Test each line individually before importing in bulk. From a terminal:
curl -x http://user:pass@host:port -s https://api.ipify.org
That should return the proxy's address, not yours. Then check the address the bot will actually talk to:
curl -x http://user:pass@host:port -o /dev/null -s -w "%{time_total}\n" https://www.retailer.com/
If those two commands disagree — the first works, the second times out — the proxy is fine and the route to that retailer is not. Run every line through the proxy checker to catch dead ones, and confirm the egress address in what is my IP so you are not comparing against a cached value. For SOCKS5, swap http:// for socks5:// — and check that the bot supports SOCKS5 at all before you buy a SOCKS5-only pool.
Latency and why it decides the checkout
A drop is a queue with a hole in it. When stock appears, every task fires at once, and the retailer's queue assigns positions by arrival time. A request that arrives 400 ms after another is behind it, and behind a few thousand other people. That 400 ms is not the shoe; it is the sold-out page.
Latency is not one number. It is DNS resolution, the TCP handshake, the TLS handshake, the request, the server's processing time and the response. With a proxy you add a hop, and the hop's distance from the retailer's servers is what matters — not its distance from you. An address in the right country but the wrong city can be 150 ms worse than one that sits in the same metro as the retailer's edge.
Three practical consequences:
- Measure to the retailer's domain, not to a generic test endpoint. The routes differ.
- Measure more than once. The average hides the tail, and the tail is what ruins a drop: a pool averaging 90 ms with occasional 1.2-second spikes will lose more checkouts than a steady 140 ms pool.
- Cut everything else. Load no images, no fonts, no tracking scripts. Every extra request is another round trip in the same race.
Routing your own traffic through a VPN on top of the proxy adds another hop and another failure point for no benefit.
What a drop costs in traffic
Less than people expect. A single drop task is a handful of page loads, a cart POST and a checkout POST — often under 5 MB if images and media are blocked, and a few megabytes more if the bot fetches product images to verify a colourway. Ten tasks on one drop is tens of megabytes, not gigabytes.
That matters because of how you are billed. Per-gigabyte billing suits monitoring, restock pinging and anything long-running and chatty. Per-address billing suits drops, because a drop is short and bursty and the scarce resource is addresses and concurrency, not bandwidth. If you buy a gigabyte plan for a single drop you will use a small fraction of it and pay for capacity you never touch; if you buy addresses you pay for exactly the thing that limits you.
The trap is the month, not the drop. Restock monitors left running for three weeks will burn through traffic quietly, and image verification on every task multiplies it. Budget the drop as small and the monitoring as the real cost.
Mistakes that lose a drop
Most failures are boring and preventable:
- Testing on the drop. The first time you run a new proxy set should not be thirty seconds before stock loads. Test hours earlier, on a different product page, with the same headers.
- Rotating mid-session. A new address halfway through means a cart tied to another IP. Session cookies and the address that minted them must stay together.
- Assuming the proxy list is live. Dead addresses do not announce themselves; they time out. Check every line before the drop, not during.
- Ignoring the ASN. An address whose ASN is a hosting provider reads as one, no matter what country it claims.
- Overloading one address. Two tasks on one address is fine until the retailer's limiter decides it is not, and then both fail at once.
- Rushing the client fingerprint. Default user agent, wrong header order, missing
Accept-Language. The proxy is blamed for a challenge the client caused. - Skipping the geography. A regional release shipped only within one country will not ship to an address in another, and stock checks may not even show the item.
- Buying on price alone. The cheapest pool is cheap because it is heavily shared and heavily used, which is the opposite of what the checkout needs.
One more, and it is not technical. Read the retailer's terms. Buying shoes is legal; circumventing a site's access controls and breaking its terms is a contract matter that gets accounts, addresses and orders cancelled, and some jurisdictions have specific statutes covering automated purchasing of limited goods. Enforcement is usually commercial rather than criminal — cancellations, bans, payment flags — but a drop won at the cost of a closed account is a loss. Provider reviews are worth reading for how a seller handles a client whose addresses burn out, because that is the moment you find out what the service actually is.
What to look for
- Latency to the retailer
- A drop is decided in the first seconds, and the round trip is most of the difference between a checkout and a sold-out page.
- How clean the pool is on retail sites
- Sneaker sites keep the most aggressive deny lists in retail, and a reused address is often already on one.
- Concurrent connections
- A bot opens many tasks at once; a capped pool turns the drop into a queue.
- City and region targeting
- Regional releases and local stock are the point, and shipping rules often follow the address.
- Billing by gigabyte or by address
- A drop is short and bursty, so per-gigabyte billing and per-address billing behave very differently.
Every provider, in order
1. NodeMaven — Best clean-IP latency
4.8 / 5 for this job · from $2.20 / GB
NodeMaven sits first because the published numbers line up with the part of a drop that cannot be recovered: the first round trip. The provider lists average residential speed under 0.2s, static ISP under 0.1s, ISP under 0.3s and mobile under 0.6s, with 95% clean IPs and a residential fraud score under 5/100 (ISP under 10/100). On a queue that opens and fills in seconds, that is the relevant claim: not peak throughput, but whether the request reaches the product page before stock disappears. The same account carries residential, mobile and ISP proxies, and sessions can be held for up to 24 hours; ISP plans are unlimited traffic with a 99.9% uptime guarantee. For sneaker work, ISP addresses are often the more useful pool for checkout steps because the address is stable through a cart, while rotating residential covers broader product and release-page requests. Country, city and ISP targeting is listed, with ZIP level claimed by the provider, so regional releases can be aimed at the right market.
The awkward part is billing. At $2.20/GB on a traffic model, a bot that opens many tasks across many SKUs burns data quickly, and high-volume scraping gets expensive. There is no free trial: the $3.50, 750 MB entry point is a paid demo, and a third-party write-up says the free-trial route is limited to registered companies that pass KYC. Refunds are pro-rated. Smaller locations are reported to be uneven in depth. The mobile pool is 250K by the provider's count and 295K+ in third-party records, so treat the larger figure with care. This card is about clean-IP latency, not cheap volume.
Key features
- Published average residential speed under 0.2s; static ISP under 0.1s
- 95% clean IPs and residential fraud score under 5/100 (ISP under 10/100)
- Unlimited concurrent sessions, sticky sessions up to 24 hours
- Country, city and ISP targeting; ZIP level claimed by provider
- ISP plans: unlimited traffic, 99.9% uptime guarantee
- No free trial; $3.50, 750 MB paid demo
Best suited for: drop tasks where first-round-trip latency and clean residential or ISP addresses matter more than per-gigabyte cost
2. Decodo (ex Smartproxy) — Best city-level targeting
4.7 / 5 for this job · from $0.08 / GB
Smartproxy — officially Decodo since April 2025 — takes second because city, ZIP and ASN targeting is the strongest fit for regional sneaker releases and local stock checks. The provider publishes a 115M+ residential pool across 195+ locations, with targeting down to city, ZIP code and ASN, plus country examples such as 16M+ US IPs and 5M+ UK IPs. For a drop that is limited to one city or one store's catchment, that is more useful than a country-level pool: shipping rules and local stock often follow the address, and the checkout can be aimed accordingly. Sticky sessions last up to 30 minutes, which is long enough to hold a cart through a checkout flow, and HTTP(S) and SOCKS5 are both supported. The published residential response time is under 0.5s and the residential success rate is 99.92%. The general response-time figure is under 0.2s, but the residential figure is the one to use for this job.
The catch is the commercial terms. The 3-day, 100 MB trial and the 14-day money-back option are mutually exclusive; using the trial removes refund eligibility. Refunds are first-purchase only, must be requested within 14 calendar days, and are capped at under 20% of a residential plan's traffic or under 10% of an Enterprise plan's. Pay-as-you-go at $4/GB is wallet-only and cannot be combined with a monthly subscription. The published fraud-score comparison does not name the scoring service. City targeting is the reason it is here; the refund rules are the reason to read before buying.
Key features
- 115M+ residential IPs across 195+ locations
- City, ZIP code and ASN-level targeting
- Residential response time under 0.5s; success rate 99.92%
- Sticky sessions up to 30 minutes; HTTP(S) and SOCKS5
- 3-day 100 MB trial excludes 14-day money-back eligibility
- Pay-as-you-go $4/GB is wallet-only
Best suited for: regional sneaker releases and local stock checks that need city, ZIP or ASN targeting rather than raw pool size
3. Novada — Best mobile carrier spread
4.6 / 5 for this job · from $0.50 / GB
Novada sits third because its mobile inventory is the widest carrier spread in this list: 2M+ 3G/4G/5G mobile IPs across 120+ carriers, alongside 100M+ residential IPs in 195+ countries with city-level targeting. On sneaker sites, mobile addresses are often treated differently from datacentre and even residential ranges, and a carrier-level pool is what lets a task look like a phone on a real network. The residential side has sessions from 1 to 120 minutes and unlimited sessions; rotating ISP has 360-minute sticky sessions across 100+ countries. Published performance is 99.99% success and sub-0.5s response, with 99.9% uptime. Billing is per GB, with mobile starting at $1.5/GB and residential at $1.3/GB; a $10 credit and a 10 GB new-user residential pack for $10 lower the entry cost.
The awkward parts matter for a drop. The protocol list is HTTP(S) only — no SOCKS5 — so some bot and antidetect setups will not connect without a workaround. The 99.99% success figure is a headline; the provider separately quotes 99% for rotating datacenter, so do not read the top number as applying to every pool. Unlimited Proxies, IPv6 and Datasets have no published price. The free trial is limited to the Scraper API; the general offer is a $10 credit, not a free trial. Mobile carrier spread is the argument. A drop that needs SOCKS5 or a published mobile latency figure should look at another card.
Key features
- 2M+ mobile IPs across 120+ carriers
- 100M+ residential IPs in 195+ countries with city-level targeting
- Residential sessions 1–120 minutes; rotating ISP 360-minute sticky sessions
- HTTP(S) only; no SOCKS5
- Mobile from $1.5/GB; residential from $1.3/GB
- $10 credit and 10 GB new-user residential pack for $10
Best suited for: mobile-carrier-heavy sneaker tasks where a real 3G/4G/5G address matters more than SOCKS5 support
4. Webshare.io — Best cheap city targeting
4.5 / 5 for this job · from $0.02 / month
Webshare.io is fourth because it is the cheapest route to city-level rotating residential in this table. The provider lists 80M+ residential IPs with country, state, city, ZIP and ASN filtering, and a rotating residential plan from $3.50/month. The per-IP side is just as blunt: proxy prices fall from $0.0299 to $0.0179 per proxy as volume grows, yearly billing takes 30% off, and there is a permanent free plan with 10 proxies and no card. For sneaker work, that makes it a sensible pool for low-cost product-page checks, stock monitoring and city-targeted requests where a single failed request is not the whole drop. Static residential (ISP) is a separate product from $6.00/month, with HTTP and SOCKS5 on the same credentials. Thread limits run from 500 to 3,000, and bandwidth tiers from 250GB to unlimited.
The catch is that this is a value pool, not a latency-first one. No response-time figure appears in the record, so there is no published number to compare with the under-0.2s or under-0.5s claims elsewhere. The unlimited bandwidth claim carries an asterisk pointing to a Fair Usage Policy. The provider's own pages give both 195 countries and 50+ locations, which makes the real footprint hard to pin down, and the static ISP list names Comcast and AT&T and then says '& more', so a specific carrier cannot be checked before buying. The 99.7% success rate and 99.97% uptime are provider figures with no independent verification in the record. Cheap city targeting is the reason it is here; a drop decided in the first second is not its strongest use.
Key features
- 80M+ residential IPs; city, state, ZIP and ASN targeting
- Rotating residential from $3.50/month; free plan with 10 proxies
- Per-IP pricing falls from $0.0299 to $0.0179 as volume grows
- Thread limits 500–3,000; bandwidth 250GB–unlimited
- Unlimited bandwidth subject to Fair Usage Policy
- No published response-time figure
Best suited for: low-cost city-targeted stock checks and product-page monitoring where budget matters more than first-second latency
5. ProxyEmpire — Large mobile pool, rollover
4.4 / 5 for this job · from $0.35 / GB
ProxyEmpire is fifth because it pairs a large mobile pool with rollover bandwidth, which suits the stop-start rhythm of a sneaker calendar. The provider lists 4M+ rotating mobile IPs across 170+ countries, alongside 30M+ residential IPs in 170+ countries. On GB-based plans, unused gigabytes roll over and never expire, so a quiet week does not write off the data bought for the next release. Sessions and concurrent connections are unlimited for approved use cases. Residential targeting goes down to country, region, city, ISP, ASN, OS/TCP and ZIP; mobile targeting drops to country, city and carrier. Rotating mobile starts at $4.50/GB, and a public API, endpoint generator and antidetect-browser integrations are listed.
The awkward part is that mobile coverage is not as deep as the headline suggests. Dedicated mobile reaches only 5+ countries and 50+ locations, and it drops city-level targeting to country, city and carrier. Datacenter is 62 countries and static ISP 24 countries, so the 170+ figure is residential and mobile, not every product. Unlimited bandwidth is priced per day or week from $200/day with no rollover, which is a different purchase from the GB plans. The EMPIRE50 discount does not apply to static, unlimited bandwidth or other flat-rate plans. The $1.97 trial excludes static ISP and datacenter and covers unlimited residential only on request. No latency or clean-IP metric is published beyond 'fast average response times'. Rollover and mobile pool size are the argument; a latency-critical drop should weigh the missing figures.
Key features
- 4M+ rotating mobile IPs across 170+ countries
- 30M+ residential IPs; ZIP, city, ISP, ASN and OS/TCP targeting
- Rollover bandwidth on GB plans; unused GB never expire
- Unlimited sessions and concurrent connections for approved use cases
- Rotating mobile from $4.50/GB; static ISP from $3.50/IP
- No published latency or clean-IP metric
Best suited for: sneaker work spread across a release calendar where unused mobile data needs to survive between drops
6. AnyIP.io — Best long sticky sessions
4.3 / 5 for this job · from $2 / GB
AnyIP.io is sixth because its standout feature — sticky sessions up to seven days — is useful for account and cart continuity but not for the opening seconds of a drop. The provider runs residential and mobile proxies on one plan at one price, metering only bandwidth, with unlimited IP rotation and unlimited concurrent requests. Sticky sessions can last up to seven days, Smart Sessions keep the IP inside the same carrier and area, and filtering covers country, ISP, region and precise GPS coordinates. IP collision prevention, change-IP-by-link, quota alerts and rollover of unused data are included. For a sneaker workflow that needs to hold a session across a release, a queue and a checkout, that persistence is the reason to consider it.
The catch is the price and footprint. The advertised $2/GB is the Enterprise rate on a 500GB plan; extra traffic is $5/GB on Micro and $4/GB on Basic, and the Micro plan is $25 for 5GB. There is no free trial — Micro is the paid trial. The country list is 38+ countries, which is short for regional releases. Published response time is around 0.6s and success rate around 98.6%, both slower than the latency-first cards. Team seats are capped at 2, 5 and 10 before Enterprise. Crypto payments sit outside the refund policy, while the 14-day money-back guarantee covers card purchases only. Long sessions are the argument; if the drop is won by the first request rather than session continuity, this is the wrong end of the table.
Key features
- Residential and mobile proxies on one plan, bandwidth metered only
- Sticky sessions up to 7 days; Smart Sessions keep carrier and area
- Unlimited IP rotation and unlimited concurrent requests
- Filtering by country, ISP, region and precise GPS coordinates
- Extra traffic $5/GB on Micro and $4/GB on Basic; $2/GB is Enterprise
- 38+ countries; published response time around 0.6s
Best suited for: long-session account and checkout workflows that need the same mobile or residential address held for days
7. ProxySale.com — Best per-day unlimited
4.2 / 5 for this job · from $0.06 / GB
Proxy-Sale.com sits seventh because its unlimited residential plan is sold by the day, and a drop is a day-shaped problem. At $69.68 per day the plan places no limit on IPs or traffic, so a task list can be as long as the machine allows without a gigabyte counter running in the background. Rotation is both random and sticky, which is what a cart session needs if it has to survive from product page through to checkout on the same address.
The pool is 86M+ addresses in 200+ countries by the provider's own count, with country, state and city targeting, HTTP, HTTPS and SOCKS5, and declared figures of under 0.6s response time and a 99.5% average success rate on residential. City targeting is the part that matters for regional releases: shipping rules and local stock often follow the address, not the country.
The catch is the billing shape. Per-day only pays off if the drop fills the day or you have other work to push through the same window; a one-hour release makes $69.68 an expensive hour, and a job that runs across several days is billed for each of them. Uptime, success rate and response time are provider-declared with no third-party audit shown. The service is not available in mainland China by the provider's own policy, and the free trial covers the Web Scraper API only, so the first residential traffic you send is paid traffic.
Key features
- Unlimited Residential with no IP or traffic cap
- $69.68 per day billing
- 86M+ residential IPs across 200+ countries, per the provider
- Country, state and city targeting on residential
- Random and sticky rotation on the unlimited plan
- HTTP, HTTPS and SOCKS5 on residential
Best suited for: a single-day drop where unlimited traffic and concurrency matter more than the flat daily price
8. FlyProxy — Unlimited sessions, city targeting
4.1 / 5 for this job · from $0.90 / GB
FlyProxy sits eighth on the strength of one line in its record: unlimited concurrent sessions, stated by both the provider and a third-party listing. On a drop, the cap that hurts first is usually the connection cap, because a bot farm opening hundreds of tasks at once turns a capped pool into a queue and a queue into a sold-out page. Unlimited sessions removes that particular ceiling.
Targeting reaches state and city level in the third-party listing, over 195 countries and regions. Third-party country tests put success at 98.7% in Germany with a 0.52s response time, 98.4% in the United States at 0.65s, and softer in smaller markets — 96.8% in Mexico at 1.00s and 97.2% in Brazil at 0.80s. Where a release is regional and the whole plan hangs on one country's stock, that spread is worth reading before you commit. Residential pricing reaches $0.90 per GB at the 10,240 GB tier, with 20 GB at $2.50 per GB.
The awkward parts are real. There are no mobile proxies in the line-up, flagged by third-party reviewers, so the address type that usually survives longest on retail sites is not on the menu. The provider states no refunds once products are sold. Pool size, protocol list and dashboard availability disagree between the provider's pages and the third-party listing — 100M+ against 69 million addresses — so buy on the version you can verify. IP lifetime is around 24 hours in third-party data, which is fine for a drop and short for anything sticky, and no free trial appears in the third-party listing even though the provider advertises 500M of traffic through sales.
Key features
- Unlimited concurrent sessions, per provider and third-party listing
- 195+ countries and regions with state and city targeting
- Third-party country tests from 96.8% (Mexico) to 98.7% (Germany)
- Response times from 0.52s to 1.00s in third-party tests
- $0.90 per GB at the 10,240 GB residential tier
- IP lifetime around 24 hours in third-party data
Best suited for: high-concurrency drop tasks in European and North American markets where city targeting and session limits matter more than a mobile option
9. Proxy-Seller — Low-cost fast residential
4.0 / 5 for this job · from $0.30 / GB
Proxy-Seller sits ninth as the low-cost fast option, and for a drop that is a straightforward trade. Residential entry is $0.30 per GB, which changes what a burst costs: mass page requests across many retailer URLs are cheap enough that you stop rationing them. The pool is 47M+ rotating IPs across 220+ locations by the provider's count, with HTTP, HTTPS and SOCKS5, a RESTful API and IP whitelisting.
The declared numbers are the sort a drop wants — around 0.35s residential response time and a 99.5%+ connection success rate — though the provider is the only source for both. Six proxy types sit under one account with no monthly lock-ins stated, so a residential plan does not commit you to a single address type for the season. Certification is unusually complete for the price band: ISO/IEC 27001:2022, ISO 9001:2015, GDPR, CCPA and ePrivacy compliance, DPA/SCC readiness, and a 24/7 support line with an advertised 30-second median reply.
The catch is mobile. Mobile coverage reaches only 20+ countries against 220+ residential locations, so if the release is somewhere the residential pool is thin, the mobile fallback is not there in the same depth. Success rate, uptime and response time are the provider's own with no independent test cited, which for a drop is exactly the moment you would want one — a 99.5% success rate still means failed tasks in the first seconds if the retailer's deny list is aggressive. Treat the figures as a starting position and keep a second provider in the config.
Key features
- Residential from $0.30 per GB
- 47M+ rotating IPs across 220+ locations, per the provider
- HTTP, HTTPS and SOCKS5 with RESTful API and IP whitelisting
- Declared ~0.35s residential response time
- Mobile coverage in 20+ countries only
- ISO/IEC 27001:2022 and ISO 9001:2015 certified, per the provider
Best suited for: cheap burst traffic on drop day where per-gigabyte cost decides the budget and a mobile fallback is not required
10. ASocks.com — Best per-address billing
3.9 / 5 for this job · from $5 / proxy
ASocks.com sits tenth because it bills per address rather than per gigabyte: $5 per proxy per month for residential and corporate, $15 per proxy per month for mobile, with volume discounts of 5% from 10 proxies, 10% from 50 and 15% from 100. For a drop, that shape suits a different job from the rest of this list. It is the account-holding case — a small number of sessions that need a stable address in one region across weeks, not hundreds of one-shot tasks in ninety seconds.
The pool is stated at 7M+ addresses across 150+ locations, which is modest against the rest of the table, and the provider says it runs its own infrastructure and links that to cleaner addresses and fewer bans. Connection success is given as 99.7%, support as 24/7 with an average response under 10 minutes, and clients as 100,000+. Because nothing is metered, a stuck retry loop or an overnight unattended watch does not add to the bill, which is the honest argument for per-address pricing here.
Every headline — 99.7% success, 150+ locations, 7M+ IPs, the G2 ranking — is provider-reported and unaudited. A 7M+ pool is small for the concurrency a release-day bot generates, so the same fixed cost that helps on a quiet account can put tasks in a queue at 10am. Mobile is three times the residential rate at $15 per month, and the record carries no city or ZIP-level targeting claim to pin a regional release to. Use it where the address is the product, not where volume is.
Key features
- Residential and corporate at $5 per proxy per month
- Mobile at $15 per proxy per month
- Volume discounts of 5% at 10, 10% at 50, 15% at 100 proxies
- 150+ locations and 7M+ IPs, per the provider
- 99.7% connection success, provider-reported
- 24/7 support with a stated average response under 10 minutes
Best suited for: a handful of long-lived sessions on one retailer where a fixed monthly cost per address beats a metered pool
11. Oxylabs.io — Large pool, ISP options
3.8 / 5 for this job · from $0.40 / GB
Oxylabs sits eleventh as the large-pool option with ISP addresses alongside it. The declared pool is 177M+ IPs in total, 175M+ of them residential, across 195 locations — enough headroom that per-task address reuse is not the constraint on a drop. The part that earns the place for sneakers is ISP: sticky datacentre-hosted addresses with the ASN of a consumer ISP, priced at $1.5 per IP on Starter and $3 per IP for dedicated, so a session can hold one fast address through cart and checkout instead of rotating mid-flow.
Rates are published for every product and every tier, so a budget is a calculation rather than a sales call: residential runs $6 per GB on Starter down to $2.5 per GB at enterprise, and datacentre per GB runs $0.60 to $0.40. Certifications are named per product — ISO/IEC 27001:2022 on Datacenter Proxies, Residential Proxies and Web API; SOC 2 Type II on the Scrape endpoint and Web Unblocker — which is more precise than most and worth reading precisely.
The catch is cost at the low end. $6 per GB on Starter is expensive for bursty drop traffic where most requests are HTML, the free entry is $5 in welcome credits, under a gigabyte at that rate, and the cheapest paid plan is $30 per month with enterprise from $2,500 through sales. VAT may apply on Starter, Basic and Flexible, so the headline rate is not the invoice. Certification scope is per product, not company-wide.
Key features
- 175M+ residential IPs across 195 locations, per the provider
- Residential from $6/GB on Starter to $2.5/GB at enterprise
- ISP proxies from $1.5/IP; dedicated ISP at $3/IP
- Cheapest paid plan $30/month; enterprise from $2,500/month
- $5 welcome credits as the only free entry
- ISO/IEC 27001:2022 and SOC 2 Type II on named products
Best suited for: drop work that wants ISP addresses and a very large residential pool on one invoice, with the budget to carry Starter rates
12. GonzoProxy — Sticky ZIP targeting
3.7 / 5 for this job · from $2 / GB
GonzoProxy sits twelfth for one specific reason: ZIP code targeting. Releases and local stock are frequently tied to a postal area, and shipping eligibility follows the address rather than the country. A pool that filters by country, city, state, continent, ZIP code and ASN lets a task list be pointed at the stores that actually hold the stock. Sticky sessions run up to 72 hours by the provider's claim, which covers a queue, a cart and a checkout without the address changing under you.
The rest of the shape suits a drop: unlimited concurrent connections, traffic that never expires and rolls into the next plan, rotating and sticky sessions, $2 per GB headline pricing, and SOCKS4, SOCKS5, HTTP and HTTPS in third-party data. The pool is 20M+ real addresses across 190+ countries. There is no KYC and access is described as live in about 30 seconds, which matters when a release is announced with short notice. Third-party country tests sit at 99.1% for the United Kingdom and 99.3% for Germany, dropping to 96.9% for Australia at 1.27s.
The catches are the ones that bite on drop day. Mobile proxies cover Poland and Ukraine only, so the address type that survives longest on retail sites is not available for most releases. The demo is not free — $1.99 for 0.3 GB, with a separate 1 GB trial row at $4.50. Two published residential plan ladders quote different per-GB rates for the same volumes, so check which one applies to your cart. The dashboard lacks detailed per-request logs, which is exactly what you want when a checkout fails at the wrong moment, and the speed and success figures are the provider's own.
Key features
- Targeting by country, city, state, continent, ZIP code and ASN
- Sticky sessions up to 72 hours
- Unlimited concurrent connections
- Traffic never expires and rolls over to the next plan
- 20M+ IPs across 190+ countries, per the provider
- Mobile proxies in Poland and Ukraine only
Best suited for: regional releases where ZIP-level targeting decides who can buy, and unused traffic can sit between drops
13. Floppydata — Cheap clean rotating pool
3.6 / 5 for this job · from $0.63 / GB
Floppydata is at thirteen because it is the cheapest credible rotating pool here, and a drop is short and bursty enough that per-gigabyte billing works in its favour. Rotating traffic starts at $1.00/GB across residential, mobile and datacenter under one account, with open volume tiers at 20% off from 96 GB and 30% off from 120 GB. Datacenter at the 120 GB tier works out at $0.63/GB, and static datacenter IPs start at $0.79 per IP per month in bundles of 20, 50 or 100.
On targeting it publishes country, city, ASN and IP-type targeting across 190+ locations per the provider, against 195 countries in third-party records — enough to place a task in the right city for a regional release, provided the city resolves the way you expect. SOCKS5, HTTP and HTTPS are all supported, there is a client API at api.floppydata.net authenticated with an X-Api-Key header, and the free trial gives 100 MB for each of residential, datacenter and mobile.
The awkward part is that "clean" rests on the provider's own 95% IP clean rate figure, and nothing independent sits behind it. The 99.99% uptime and sub-0.3-second response claims are the same kind of number. The 100 MB trial per type is a connectivity check, not a load test, so it will not tell you how the pool behaves against a retailer's deny list under volume. Documentation is thin for advanced setups such as custom rotation logic, and there is no phone support if a drop goes wrong at an awkward hour.
Key features
- Rotating traffic from $1.00/GB across residential, mobile and datacenter
- 20% off at 96 GB, 30% off at 120 GB as open tiers
- Static datacenter IPs from $0.79 per IP per month in bundles of 20, 50 or 100
- Country, city, ASN and IP-type targeting
- SOCKS5, HTTP and HTTPS
- 100 MB free trial for each of residential, datacenter and mobile
Best suited for: buyers running short, bursty drop traffic who want per-gigabyte billing and city targeting at a low entry price
14. Nsocks — Broad city targeting
3.4 / 5 for this job
Nsocks is here for one reason: city targeting at a breadth most of this list does not publish. The provider quotes country and city targeting across 190+ country and city-level regions, with country, city, ISP and ASN targeting reported by third parties. For a regional release — a Europe-only drop, or a store that ships inside one country — that is the lever that decides whether a task looks like it is shopping locally or tunnelling in from somewhere else.
The pool is quoted at 80M+ residential IPs by third parties, though the provider's own home page carries both 80M+ and 90M+ in different places, which is worth knowing before sizing a plan on the figure. Residential is billed per gigabyte across tiers running from 10 GB to 500 GB, with static residential and datacenter sold per IP on fixed monthly billing, unlimited bandwidth and a 20 Mbps cap on the static residential plans. Protocols are HTTP, HTTPS and SOCKS5, and there are listed integrations with the main anti-detect browsers plus Playwright, Selenium and an API.
The catch is that there is no free trial and the provider states it does not refund products sold, so pool quality on a retail site is something you find out after paying. Third-party records put the founding date at 26 August 2024, a short history for a job where a pool's reputation on the big sneaker sites is built over years. Per-gigabyte rates for the residential tiers are not printed on the pricing page, so budget from the tier structure rather than a quoted rate. Support is 24/7 by live chat and email, which covers most of what a drop needs operationally.
Key features
- 80M+ residential IPs claimed by the provider, with 90M+ also quoted on its home page
- Country, city, ISP and ASN targeting
- 190+ country and city-level regions, per the provider
- Residential billed per GB across tiers from 10 GB to 500 GB
- Static residential and datacenter sold per IP on fixed monthly billing
- HTTP, HTTPS and SOCKS5
Best suited for: buyers chasing regional releases who need city-level targeting and will accept a no-trial, no-refund purchase
15. MobileProxy.Space — Unlimited mobile traffic
3.3 / 5 for this job · from ₽490 / day
MobileProxy.Space is the mobile-traffic option here: plans bill by rental period with unlimited traffic rather than per gigabyte, with a minimum rental of one day from 490₽/day. Mobile addresses sit behind carrier NAT shared with real handsets, which is why they survive retail deny lists longer than datacenter ranges. A per-day unlimited plan can therefore work out cheaper than metering gigabytes through a pool that keeps getting flagged.
Concurrency is stated at up to 250 simultaneous connections per mobile proxy, and rotation runs on a timer or by hitting an API link, which suits a bot that wants a fresh address mid-drop. There is an API and an MCP server for Claude Code, Cursor, Python, Node.js, PHP and cURL, and dead proxies are replaced automatically according to the provider. Geography is the part to check: the provider quotes 258 cities and 53 countries, while third-party records give 30+ countries and 130+ mobile operators. Those sets do not agree and the narrower one is the safe assumption.
The awkward part is latency. Declared figures give an average ping of 348 ms and a response time of 3.24 seconds, with download speed at 10.00 Mbps. On a drop decided in the first seconds, that round trip is most of the gap between a checkout and a sold-out page. Third parties also flag the prices as relatively high, and there are reports of laggy proxies and late-night maintenance windows. It suits mobile-only tasks and short campaigns where unlimited traffic matters more than a fast first byte.
Key features
- Mobile plans billed by rental period with unlimited traffic, from 490₽/day
- Minimum rental of one day
- Up to 250 simultaneous connections per mobile proxy
- IP rotation on a timer or by API link
- 258 cities and 53 countries quoted by the provider; 30+ countries and 130+ operators in third-party records
- Declared average ping 348 ms and response time 3.24 seconds
Best suited for: mobile-only tasks and short unlimited-traffic campaigns where per-day billing beats metering gigabytes
16. Live Proxies — Unlimited threads, private IPs
3.1 / 5 for this job · from $2
Live Proxies earns its place on concurrency. The provider states no concurrency limit, unlimited threads and unlimited requests, which is the line that matters when a bot opens hundreds of tasks in the same second; a capped pool turns a drop into a queue. IPs are allocated privately, which the provider says are unique and shared only with you — the opposite of a shared pool where an address may already be on a retail deny list because of someone else's session.
Sticky sessions run 60 minutes on rotating residential and 30 on rotating mobile, unused bandwidth rolls over to the standby plan, and dashboard access, API access and IP whitelisting are included. Activation is quoted at around 10 minutes, up to 20. Mobile rotating plans start at $55.00/plan, and B2C residential plans start at $70.00 for 200 IPs and 4 GB over 30 days.
The catch is coherence. The home page prices static residential at $65.00/plan while the pricing page marks it permanently unavailable, and the $45.00, $55.00 and $65.00 starting prices do not reconcile with "B2C plans starting from $70". The free trial goes to custom plan users after an agreement, with B2C buyers told to purchase a small plan instead. SOCKS5 is listed for B2B only, so a consumer stack that needs it is out of scope. Geography is described as countries or state-level options with no count, and state IPs are US only — thin for a job where you need to know which city an address resolves to before a regional release.
Key features
- No concurrency limit, unlimited threads and unlimited requests, per the provider
- Private IP allocation, unique to the account
- 60-minute sticky sessions on rotating residential, 30 minutes on rotating mobile
- Unused bandwidth rolls over to the standby plan
- SOCKS5 on the B2B tier only
- Rotating mobile from $55.00/plan; B2C residential plans from $70.00
Best suited for: high-concurrency setups that need private addresses and will verify location themselves before a drop
17. ChangeMyIP.com — Cheap shared testing
2.2 / 5 for this job · from $1
ChangeMyIP.com closes the list because the product is shared proxies at $1.00 a unit, and shared is the thing a sneaker drop punishes hardest. The provider says plainly that shared proxies share bandwidth and put multiple users on the same IP address, and that performance can suffer on busy servers. On retail sites with the most aggressive deny lists, an address several strangers have already used is often flagged before your first request goes out.
What it is good for is the cheap end of testing a stack: instant activation, unlimited bandwidth on the plans it sells, 1Gb/s ports, multiple IP subnets, no contract, and a protocol range well beyond HTTP — SOCKS5, ShadowSocks, Trojan, DNS proxy and Wireguard VPN, with crypto payments through Zonda alongside cards and PayPal. Dedicated HTTP is $1.80, shared HTTP and shared SOCKS5 $1.00, SOCKS5 and ShadowSocks $2.20, Trojan $3.00, DNS proxy $2.00 and Wireguard $4.00.
The catches stack up for this job. A $10 minimum order applies even if you want a single $1.00 proxy, torrents and paypal.com are blocked at proxy level on every product, and the 30-day money-back guarantee does not apply once the service has been used, with refunds limited to bugs or slow performance. The "fully anonymous and undetectable" line is unverified marketing. Shared datacenter addresses are the wrong tool for a limited release — buy here to confirm a client speaks the protocol and handles authentication, not to win stock.
Key features
- Shared HTTP and shared SOCKS5 from $1.00 per unit
- Dedicated HTTP at $1.80; SOCKS5 and ShadowSocks at $2.20
- Unlimited bandwidth with 1Gb/s ports
- $10 minimum order
- HTTP, SOCKS5, ShadowSocks, Trojan, DNS proxy and Wireguard VPN
- 30-day guarantee void once the service has been used
Best suited for: protocol and configuration testing on a tight budget, rather than contested limited releases
Sneakers proxies: questions we get
Are free proxies any use for a sneaker drop?
Not for the checkout. Free lists are public by definition, so an address has been used by anyone who found it, often on the same retail sites, and many are already on deny lists before you load them. They are useful for two things: learning the import format, and seeing how quickly a retailer challenges a bad address.
Can a retailer tell I am using a proxy?
Sometimes directly, always indirectly. A hosting ASN or a mismatched timezone and language header gives it away immediately. Even a clean residential address can be inferred from behaviour — request rates, TLS fingerprints, header order and cookie patterns that no browser produces.
Do I need a separate proxy for every account?
If those accounts are used on the same site, yes, and permanently — not just for the drop. A retailer that sees two accounts sharing an address links them, and it links the payment methods and shipping addresses attached to them. Sharing an address between accounts is how people lose whole account portfolios rather than one order.
What is the difference between rotating and sticky sessions?
A rotating proxy gives you a different exit address on a schedule or per request. A sticky session holds one address for a set period, usually minutes. Drops need sticky — the cart and the session cookie belong to the address that created them, and rotation mid-flow empties the cart.
Will using a proxy get my order cancelled?
It can, and proxies are not the usual reason. Orders are more often cancelled for breaching purchase limits, mismatched billing and shipping details, or a payment method flagged by the retailer's fraud checks. An address with a poor history on that specific site is a contributing factor.
Is sneaker botting legal?
Buying shoes is legal. Using a bot generally breaches the retailer's terms of service, which is a contract issue rather than a criminal one — the usual outcome is cancelled orders and closed accounts. Some jurisdictions have specific laws aimed at circumventing access controls on limited-availability goods, so the answer depends on where you are and what you are buying.
Why did my proxy work in the checker but fail on the drop?
The checker tests the proxy against a neutral endpoint, which tells you the credential and the route are alive. It says nothing about how that address is regarded by the retailer's bot protection, or whether the route to that particular domain is slow. Always test against the retailer's own domain.
Do I need residential proxies for a Shopify site?
Shopify storefronts are usually lighter on bot protection than the major retailers, so less is needed than people assume. That said, a drop on a high-traffic store may still front onto a queue or a challenge, and a shared datacenter address will be the first thing caught. Match the address type to the protection, not to the shoe's resale value.
How do I check whether a proxy is already banned on a retailer?
Load the product page through the proxy in a browser or with a plain request and look at the status and the body — a 403, a redirect to a challenge page or an unusually small response all indicate a block. Do this before the drop, on a page you do not care about. A block on a product page usually means a block on checkout.
What is an ASN and why does it matter for sneaker proxies?
An autonomous system number identifies the network an address belongs to. Retailers keep lists of networks associated with hosting and with known abuse, and will challenge or block whole ranges on that basis. Two addresses in the same country can behave completely differently because of the network behind them.
Should I use a VPN instead of a proxy for a drop?
No. A VPN puts your traffic through one address and adds a hop, which is the opposite of what a drop needs — many addresses and the shortest path. It also gives every task the same egress, so one block takes everything down. VPNs do not give you per-task addressing or city targeting.
How far ahead of a drop should proxies be set up?
Days, not minutes. You want time to test every line, confirm the retailer does not challenge it, measure latency from the address to the site, and check that the concurrency limit covers your task count. Anything bought or configured an hour before a drop is untested, and untested means unknown.
Other jobs people buy proxies for
- Web scraping
- SEO and rank tracking
- Ad verification
- Price monitoring
- Multi-account management
- Antidetect browsers
Some links on this page are affiliate links: if you buy through them we may earn a commission, at no cost to you. It does not change the order above — see our monetisation policy.