Protection
Stop the attacks, not your users
ZWAF is a web application firewall for your websites and APIs, with bot and L7 DoS protection on the same appliance. Injection attempts, stolen passwords tried one after another, scrapers and floods reach every public application, and ZWAF stops them in front of your servers. You start in a mode that only reports, tune the policy on real traffic and switch to blocking when you are confident.
WAF and API security
A WAF policy that fits each application
Generic rules either miss attacks or block real users. ZWAF combines known attack signatures with a model of what your application really accepts. Start from a ready-made policy, let ZWAF learn from real traffic, and decide yourself what goes live.
Signatures kept current for you
ZWAF Labs, at Zero Day Technologies, maintains a signed signature set: the OWASP Core Rule Set rewritten for ZWAF, plus our own rules. Tell ZWAF which databases, systems, languages and web servers you use, and only their signatures run. By default, new signatures report for seven days before they block.
Only what your application accepts
Allowed URLs, methods, parameters with their types and lengths, headers and cookies. Hidden form fields are read-only and cookies are signed, so any tampering shows up as a violation.
APIs held to their contract
Import OpenAPI 3 or Swagger 2, JSON Schema or XSD, and requests that do not match are stopped. GraphQL gets limits on depth, complexity, aliases and batches, with introspection blocked.
Learns from your real traffic
The policy builder watches the traffic and proposes URLs, parameters and their types, each with a confidence score. Requests flagged as attacks are left out of what it learns. Accept proposals one by one, or let those above a threshold apply themselves.
Report first, block when you are ready
Run a policy in a mode that only reports, then switch it to blocking. Blocked users see a support ID they can quote to you, and a false alarm becomes an exception in one step, straight from the event.
Cover scanner findings before the code is fixed
Import a report from OWASP ZAP, Burp Suite, a SARIF file or other common scanners, and ZWAF drafts a policy that blocks each finding. ZWAF marks a finding as covered once it has checked that the draft blocks it.
Bot protection
Stop bad bots without making people solve puzzles
Most visitors never see a challenge. ZWAF sorts each client from signals it already has, and asks only when it is not sure. Search engines keep indexing your site. Scrapers and password-guessing tools are recognised and stopped.
Good bots in, bad bots out
Real search engine crawlers, verified by DNS, keep their access. Bots that only pretend to be one, attack tools, clients that are not browsers and plain HTTP libraries are recognised, and each kind gets the action you choose.
Signals it already has
The User-Agent, the order of the headers, the JA4 TLS fingerprint, the reputation of the address and, on websites, a small script that reports headless browsers and automation tools.
No outside captcha service
A JavaScript proof of work that needs no click, or ZWAF's own CAPTCHA, drawn on the appliance, with a version for screen readers in English or Romanian. No third party sees your users.
Mobile apps covered too
The ZWAF SDK for Android and iOS gives each installation its own key, registered with Play Integrity or App Attest, and signs every request. Replayed or altered requests are treated as automated.
Account takeover attempts stopped
Failed logins are counted per address, per user, per device and overall. Above the limits a source is stopped or must prove it is a browser, and each attempt can be checked against known leaked credentials.
Scrapers and abused business flows
Hidden trap links and navigation patterns give scrapers away. Flows such as checkout can be declared step by step, so their order, timing and device are enforced and no step can be replayed.
L7 DoS protection
It knows your normal traffic, so it spots the flood
A fixed threshold reacts too early or too late. ZWAF learns what is normal for each application and each URL: the request rate, your servers' response time and their share of errors. An attack shows up as a departure from that, and ZWAF answers in steps.
Two ways to detect
By rate, when requests per second climb above normal, for the whole site or for a single URL. By stress, when your servers slow down or start to fail.
The attackers grouped for you
During an attack ZWAF finds what the attacking requests have in common: a subnet, a country, a TLS fingerprint, a User-Agent or a URL. It limits them as one source, instead of slowing everyone down.
A measured response
A challenge first, then limits per client, device, URL and signature, and a queue in front of the URLs that cost your servers the most.
Slow attacks too
Limits on incomplete connections per address, minimum send and read rates, and a time limit for the request headers stop slowloris, slow POST and slow read.
A report for every attack
Follow an attack live: rates and response times against normal, the attack signatures and the top sources, refreshed every two seconds. When it ends, its report stays: start, end, peak and what was stopped.
Next step
See ZWAF in front of your applications
Tell us what you run and how your traffic looks. We will show you how ZWAF would protect it and answer your questions about licensing and support. Resellers and integrators are welcome too: we are building our partner network and would like to hear from you.