How to detect AdsPower in 2026
AdsPower is an antidetect browser: a fake browser that gives every profile its own device fingerprint, its own proxy and its own storage, so one person can run many accounts that each look like a different computer. It is sold openly, with two free profiles, to sellers who run several stores on Amazon, eBay, Shopify, Shopee and Temu, and to teams that manage ad and social media accounts. Robots, a Local API and, since September 2026, an AI Agent work those profiles for the operator.
You will not catch it by name. Maskbreak's device check flags the class of tool, a browser whose answers contradict each other: the response carries device.antidetect and the reason antidetect_browser, and the default decision is block. The script driving the profile and the proxy in front of it are separate signals.
Last reviewed: September 2026. Facts about AdsPower come from its own site and release log, read on 25 September 2026.
What AdsPower is in September 2026
AdsPower is made by SUNFLOWER TECH PTE. LTD., the company named in its site footer. It runs two engines of its own: SunBrowser, built on Chrome, and FlowerBrowser, built on Firefox. Both follow the upstream releases: the release log shows a Chrome 152 kernel on 9 September 2026, Firefox 153 support on 15 September and a Chrome 153 kernel for Windows, Mac and Linux on 23 September (release log). Chrome itself now ships every two weeks, so any version number here goes out of date quickly.
Around the engines sit the tools for running profiles in bulk. The site lists a fingerprint editor with "20+ options and 50+ customizable parameters", a proxy per profile (HTTP, HTTPS, SOCKS5 or SSH), batch import and export of profiles, a Window Synchronizer for working several windows at once, team members with roles and per-profile permissions, RPA robots for batch operations, a Local API for your own scripts, and MCP control (adspower.com). The AI Agent rolled out in September 2026 and is paid for in credits, from $3 for 200 (launch post); the 23 September release lets it run several sessions at once and generate RPA templates.
The free plan is two profiles with no card. Paid prices appear only after you choose a plan on the pricing page and change with promotions, so we do not quote them.
Much of this is ordinary agency and multi-store work, and running several accounts can break a platform's terms without breaking any law. The same kit, many profiles with a proxy each and robots to work them, is also what a fake-signup, bonus-abuse or referral-farming operation needs. That is where fraud teams meet it.
What gives an AdsPower session away
Maskbreak does not look for SunBrowser or FlowerBrowser by name, and the response never says "AdsPower". Each visit is judged on whether the claimed device holds together, whether a script is driving it, where the traffic comes from and which of your accounts the device has touched.
device.antidetect is true, the reason antidetect_browser is added and the default decision is block.device.tampering_score is a 0–1 value from the device check, not a probability of fraud. On our test bench it read 0.0029 on the clean baseline and on every spoofed profile while the fake-browser flag moved (lab notes). Route on decision and reasons; do not set a threshold on the score.device.automation is true, the reason is automation_detected and the default is block. It does not name the tool, and a careful script does not always show those signs. Bot detection covers the signal in full.emulator_detected and blocks. A virtual machine is reported as device.virtual_machine with the reason virtual_machine, and on its own it asks for review, never a block: plenty of developers browse from one.proxy_detected and blocks, a VPN on its own adds vpn_detected and goes to review, and network.service names the operator when known. An exit it does not recognise passes the network check, which is why the device flag has to stand on its own. Proxy detectionaccountId and device.linked_accounts counts how many of your accounts this device has been used on; more than one adds the reason multi_account_device, which changes the decision only if you add a rule for it. Links stay inside your account, stored as one-way hashes. device.times_seen is different: it counts sightings in the device's retained history across all Maskbreak customers (periodic cleanup removes records inactive for more than 90 days), not visits to your app. Profile tools are built to beat linking, since each profile can present a new visitor ID, so treat it as an extra layer. Multi-accounting playbookaccountId, account linking) rather than resting on one check.What the response looks like
An example response for a fake browser driven by a script from a home address that the network check does not flag, trimmed to the fields that matter. It is computed from the default weights as an illustration, not an AdsPower test result.
// example response, POST /v1/evaluate
{
"decision": "block",
"risk_score": 75,
"reasons": ["antidetect_browser", "automation_detected"],
"network": {
"vpn": false,
"proxy": false,
"datacenter": false,
"anonymous": false,
"tor": false,
"residential": true,
"service": null
},
"device": {
"antidetect": true,
"automation": true,
"emulator": false,
"virtual_machine": false,
"tampering_score": 0.0029
}
}
The score is 75: 35 for the fake-browser flag and 40 for automation. Either one alone would block, so route on decision, not on the number. residential is true only because neither the proxy nor the cloud-server flag fired; it does not prove a home connection.
Integration in one pass
Add the browser snippet to the forms where these profiles show up: signup, login, checkout, or a bonus or referral claim. Your server forwards the two tokens it collects to POST /v1/evaluate with your API key and routes on decision and reasons. Start in watch mode, which logs each decision and lets every visitor through, and switch to enforce once the Events log looks right. The quickstart has the handler and the API reference lists every field, or hand integrate.md to your coding agent.
If some of your customers run AdsPower on purpose. Agencies and multi-store sellers are AdsPower's market, so they may be yours too. Set the "Fake browser (antidetect)" rule to review on the console's Rules tab. A rule changes the decision only for the signal it names, so a proxy, Tor, automation or an emulator on the same visit still blocks. Block or review antidetect browsers? weighs the choice.
For the whole category, the antidetect browser hub covers how these browsers work and lists makers and engines. AdsPower also has a row in our 2026 antidetect tests, where it is queued and not yet run.
Free plan: 10,000 visitor checks a month, no card. Paid plans from €29 a month.