Skip to main content
Demos

See how we work

Short, interactive examples of our security and software work. Pause any demo, step through it, or read the text description.

All demos on this page are sample scenarios built with synthetic data; they involve no real customer system, scan or attack.

Simulation · Sample scenario

Edge protection during an HTTP flood

The wave of fake requests is filtered before it reaches the application, so real users keep using the site.

Traffic sourcesEdge / CDNRate limitingBot managementWAF rulesApplication● Load normal
  • User
  • Bot
  • Legitimate request
  • Attack request
  • Blocked at the edge

Step 1 of 6 · Normal traffic

Normal traffic
Text description

Real users (circles) and bot sources (diamonds) are on the left. Normally only user requests pass through the edge layer to the application, and application load stays low. In the attack scenario bots send a large number of HTTP requests and application load rises quickly. With protection on, rate limiting, bot management and WAF rules engage at the edge; bot requests are blocked at the edge (cross mark), user requests keep reaching the application and load returns to normal. With protection off, load stays high and real users' requests slow down.

Request a security assessment

Rate limiting, bot management and WAF are different controls; together they reduce application-layer (L7) load. Volumetric L3/L4 attacks are absorbed upstream in the provider's network before traffic reaches you. Figures are illustrative.

Simulation · Interactive diagram

Layered security architecture

See what each layer stops, and where the monitoring path sits.

  1. Meets traffic closest to its source and reduces application-layer load through caching, rate limiting and bot management.

A request passes through in this order; each layer targets a different risk.

Edge / CDN: Meets traffic closest to its source and reduces application-layer load through caching, rate limiting and bot management.

Text description

Traffic path: Internet, Edge/CDN, WAF and policies, network security, application/API. The monitoring and decision path is separate: edge, network and application logs are collected, analysed through correlation, and result in a notification, human review or a policy-based response. The monitoring path is not inline; it does not stop requests itself.

Simulation · Sample workflow

From code scan to a verified fix

We trace a finding to the exact line, explain the risk and check the fix.

routes/users.js
 router.get("/users", async (req, res) => {
   const email = req.query.email;
   const sql =
     "SELECT id, name FROM users WHERE email = '" + email + "'";
   const { rows } = await db.query(sql);
   res.json(rows);
 });

Step 1 of 6 · Scan

Scan
Text description

Four files are scanned and one finding appears in routes/users.js. Because the user-supplied email is concatenated into the SQL query as text, the lines are flagged and the SQL injection risk (CWE-89) is explained. The fix turns it into a parameterized query that sends the value separately; old lines are shown as removed and new lines as added. The last step shows a sample verification result; this is not a real customer scan.

Simulation · Sample workflow

Bug → test → fix → retest

We find the bug, fix it, and confirm it by rerunning the same scenario as a test.

Bug report

“When I apply a 10% discount code, the cart shows 390 instead of 360.”

cart.js
 export function cartTotal(items, discountPct = 0) {
   const sum = items.reduce((acc, it) => acc + it.price * it.qty, 0);
   return sum - discountPct;
 }
$ CART_IMPL=buggy node --testReal output
Test not run yet…

Step 1 of 6 · Bug report

Bug report
Text description

A user reports that with a 10% discount the cart shows 390 instead of 360. A test reproducing the bug is run and fails: expected 360, got 390. The root cause is that the discount percentage is subtracted from the total as if it were an amount. The line is fixed to apply the discount as a percentage, and rerunning the same test makes both tests pass. The terminal lines are real test output from the sample project in the repository.

Let's talk about your project

The flow is an animation. The terminal lines are real output (trimmed) from the demo-samples/cart-total example in the repository, run with node --test; test names are in Turkish as recorded.

Simulation · Sample scenario

Web security testing and reporting

We separate potential findings from verified vulnerabilities, prioritise by business impact and retest the fix.

  1. Target & permission
  2. Scope
  3. Scan
  4. Validation
  5. Prioritisation
  6. Remediation
  7. Retest

Authorised target

https://app.example.com

  • Written permission and test window approved
  • Contact person and emergency stop agreed
  • Only the defined scope is tested

Step 1 of 7 · Target & permission

Target & permission
Text description

Seven steps: authorised target app.example.com; in-scope and out-of-scope addresses; security checks; validation of findings (server version header is potential and low, access to another user's invoice is verified and high); prioritisation by business impact; recommended fix; a sample retest where the same request is now rejected with 403.

Request a security assessment

Sample scenario: app.example.com is a reserved example domain and no real scan was performed. Real tests are run only with written permission and within a defined scope.

Simulation · Sample scenario

SOC event flow: from log to decision

Not every alert is handled the same way: some go to human review, some to an automated policy response.

Case 1 · Off-hours admin login

  1. Event
  2. Log
  3. Correlation
  4. Priority
  5. Outcome

Event

Successful VPN login with an administrator account at 03:12.

Case 2 · Hidden file probing

  1. Event
  2. Log
  3. Correlation
  4. Priority
  5. Outcome

Event

Requests for /.env and /.git/config from a single IP address.

Step 1 of 5 · Event

Event
Text description

Two cases pass through the same five stages: event, log, correlation, priority and outcome. In case 1 an administrator login in the middle of the night does not match the user's usual profile; it gets medium priority and goes to analyst review. In case 2 a single IP address requests 40 sensitive files per minute; as a known scanning pattern it is temporarily blocked by policy and the team is notified.

Request a security assessment

Sample scenario; IP addresses are from ranges reserved for documentation. Automated monitoring and policy-based response are not the same as a SOC service run 24/7 by a human team; scope and response times are defined in the service agreement.

AI, Cybersecurity & Enterprise Software Engineering

Secure. Scale. Innovate.

We build secure, AI-powered digital platforms for ambitious companies across Turkey and Europe.