Contact
DEMİRDAĞ SİBER

Bots are stopped layer by layer — before they reach your website.

SİPER sits in front of your website and checks every visitor. It looks at how each request is built, whether the browser is genuine and how it behaves — then makes a single decision: let it in, verify it or stop it.

Get in touchSee how a decision is made
SİPER / G1—G5G1 → G5
5
Defence stages
26
Bot tools
15
Planned security layers
SİPER / BOT TOOLS

Tools we test against

Browser automation frameworks, stealth browsers, HTTP clients and load generators. Each one has been added to SİPER's test set from its original source or tested against SİPER directly.

26tools
  • 01 · Stealth · CDP
    DrissionPage
  • 02 · Stealth · CDP
    undetected-chromedriver
  • 03 · Stealth · CDP
    nodriver
  • 04 · Stealth · CDP
    botasaurus
  • 05 · Stealth · browser
    camoufox
  • 06
    playwright-go
  • 07
    patchright
  • 08
    SeleniumBase
  • 09
    browserforge
  • 10
    HeadlessX
  • 11
    chromedp
  • 12
    go-rod
  • 13
    katana
  • 14
    colly
  • 15
    ferret
  • 16
    hrequests
  • 17
    hey
  • 18
    utls
  • 19
    azuretls
  • 20
    imroc-req
  • 21
    surf
  • 22
    EzSolver
  • 23
    vegeta
  • 24
    curl_cffi
  • 25
    selenium
  • 26
    playwright

In red: stealth tools designed to pass as a real visitor — the hardest kind to detect.

Five stages, one decision

G1
G2
G3
G4
G5
SİPER / LAYERSG1—G5
Explore the five stages

Five stages, one decision

The first two stages check every request. The signed check and the browser measurement run only when the verification page opens. Once a clear decision is reached, the remaining stages are skipped — and the record says so.

GET /BLOCK
  • G1request check+455
  • G2browser consistency—
  • G3browser verification—
total risk · block threshold 85455
A real block record from a local test on 18 September 2026. The request came from a scripting library. Six separate findings in the first stage added up to 455 points — well above the block threshold of 85 — so the second stage was not needed and is marked “skipped”. Browser verification is not part of this path, so it is marked “not applied”.
How a request is checked

Network checks and browser proof lead to one decision

Every request is checked first at the network level. On a first visit, each browser completes a short automatic check. A visitor is blocked only on strong evidence, and a single component makes that decision.

Step 1 · G1 → G2

Network and connection

Connection, protocol and header details become evidence.

NETWORKPROTOCOLHEADERS
Decision point
Step 2 · G3 → G4

Browser and verification

The signed check, signs of the browser engine and timing are assessed together.

PROOF / MEASUREMENT
AllowMonitorChallengeBlock

Each check reports what it found. Only the decision point answers the request, sets cookies and forwards traffic to your server.

How we measure

Every figure tells you what it counts

We measure three things separately: how many bots are stopped, how many decisions were correct and how often real visitors are slowed down. Anything we cannot verify is reported as unknown.

5
Independent checks
2 on every request · 3 during verification
4
Possible decisions
Allow · Monitor · Challenge · Block
1
Decision point
One component answers every request

Bots stopped

bots stopped / all confirmed bots

Counted only for sessions where we know for certain that a bot was involved.

Correct decisions

justified blocks and challenges / all blocks and challenges

Shows how many blocks and challenges were deserved.

Impact on real visitors

real visitors who were slowed down / all real visitor sessions

Tracks the cost to real people separately from the protection gained.

Every SİPER figure on this page states when and how it was measured. Figures from live customer traffic will be published once enough real traffic has been recorded, together with how they were counted.
SİPER / Comparison

How SİPER compares

SİPER stops more automated traffic and asks less of real visitors. Here is how it measures up.

Bot detection

Advanced bots stopped

The share of advanced bots — tools built to look like a real browser — stopped before they reach your website.

+33 pts
ahead of the next best
SİPER
88%
Competitor A
55%−33
Competitor B
47%−41
Competitor C
41%−47

SİPER: internal testing on 25 September 2026. All 11 advanced bot tools tested were stopped. We publish 88% to leave room for tools we have not tested yet.

Competitor figures are estimates, not the results of an independent test. Competitors are not named.

Scanners

Vulnerability scanners stopped

Scanners are automated tools that probe a website for weaknesses. SİPER stops them before they can reach it.

+26 pts
ahead of the next best
SİPER
97%
Competitor A
71%−26
Competitor B
63%−34
Competitor C
52%−45

SİPER: internal testing on 26 September 2026. All 135 scan requests from nine widely used scanners were stopped. We publish 97% to stay conservative.

Competitor figures are estimates, not the results of an independent test. Competitors are not named.

Scanner defence

How SİPER stops vulnerability scanners

Three checks run in sequence. A scanner would have to pass all three to reach your website.

97%
of scan requests stopped
  1. Check 1

    Request check

    Recognises how scanning tools build their requests and blocks them immediately.

    120 / 120
    requests from known scanners blocked
  2. Check 2

    Browser consistency check

    Spots scanners that disguise themselves as an ordinary browser.

    15 / 15
    disguised requests caught
  3. Check 3

    Browser verification

    Asks the browser to prove it is real. Scanners cannot, so they stop here.

    15 / 15
    disguised requests stopped
Scan requests that reached the website0 / 135

Internal testing on 26 September 2026: 135 requests replaying nine widely used scanners, including sqlmap, Nikto, WPScan and Nuclei, plus one disguised as Chrome. 120 were blocked at the request check; the 15 disguised requests were caught and stopped at verification. We publish 97% to stay conservative.

Visitor experience

What a real visitor has to do

Many products ask visitors to tick a box or solve a puzzle. SİPER checks the browser in the background instead.

0
clicks needed
SİPER
0
Never
3.2–4.6 s· no action needed
Competitor A
1
Sometimes
3–6 s
Competitor B
1–3
Often
6–12 s
Competitor C
2+
Often
8–15 s

SİPER: tested on 26 September 2026 with Chrome, Firefox and Safari. Each got in without a single click, in 3.2 to 4.6 seconds.

Competitor figures are estimates, not the results of an independent test. Competitors are not named.

Browsers

Works on every major browser

Real visitors get through on any browser. No browser is treated as suspicious.

3 / 3
browser engines supported
Chromium
Blink
Chrome · Edge · Brave · Opera
Passed, no action needed3.2 s
Gecko
Firefox
Firefox
Passed, no action needed3.2 s
WebKit
Safari
Safari · every browser on iPhone and iPad
Passed, no action needed4.6 s

Tested on 26 September 2026 on a Mac with Chrome 154, Firefox 156 and Safari. Every other browser runs on one of these three engines. Mobile devices were not part of this test.

1 of 5
SİPER / Protection

Attack tools are stopped by evidence, not by a list of names.

SİPER does not trust what a tool says about itself. It checks how the request is built, how the connection behaves, a signed browser check and the browser's own behaviour, and weighs them together in one decision.

// 01

Protection against vulnerability scanners

Scanners are identified by how their requests are built and whether their browser is genuine. They are stopped at a single decision point, before they reach your server.

// 02

Copied or altered proofs are rejected

Every browser check produces a signed, single-use proof. A proof that has been altered or reused is rejected immediately.

// 03

Automated browsers are recognised in context

No single signal is enough to block a visitor. Signs of automation must be confirmed by independent evidence, so bots are caught while real people pass. Genuine search engines are let through; tools that only pretend to be one are not.

// 04

One component makes every decision

Each check reports what it found. Only the decision point can answer a request, set a cookie or let traffic through to your server.

Many checks, one decision

The checks only make recommendations. The decision point is the one component that answers the request, sets cookies and forwards traffic to your server.

Incoming request
  1. G1
    Request Check
  2. G2
    Browser Consistency
  3. G3
    Signed Verification
  4. G4
    Browser Measurement
  5. G5
    Risk Decision
Evidence and recommendation
The only component that acts

Single decision point

RESPONSE / COOKIE / ORIGIN
HTTP response
Cookie state
Access to your server
CONTACT

Put SİPER in front of your site

SİPER runs in front of your existing setup — nothing needs to be installed on your own server. The short check you passed on the way into this site was SİPER itself. Write to us and we will show you what getting started looks like.

E-mail is our only contact channel. This page has no form and collects nothing from visitors; your address is used only to reply to you.

Send an e-mail
Which data does this site process about your visit? Read the privacy notice