Kavra LabKavra at a glance
| Question | Answer |
|---|---|
| What is it? | A bot and fraud detection platform for websites, web apps and mobile apps (native iOS and Android SDKs), with a server-side request API. See Platform. |
| Who makes it? | Kavra Lab, a brand of SIA ADVERTIMO, registration number 42103094905, Riga, Latvia (EU). See About. |
| What does it answer? | Who is really behind each request: a real person, a verified bot or AI agent, an automation tool, or something impersonating a browser, and why. |
| How does it decide? | It analyzes 3,000+ data points per visit across network, device and environment, browser integrity, behavior, and identity and history; an AI/ML risk engine weighs them together and looks for contradictions between layers. |
| What do visitors see? | Nothing. Every check runs in the background; Kavra never shows a CAPTCHA, puzzle, slider or checkbox. Any extra step, such as a one-time code, is the customer's own choice. |
| What does the customer receive? | A plain-language headline, the findings behind it, risk levels by domain, network context and a recommendation: allow, verify or block. |
| Who decides the outcome? | The customer's backend. Kavra never blocks users on its own. An observe-only mode shows results before anything is enforced. |
| How is it integrated? | One script on the page (under 64 KB, asynchronous) and one server-side call; no DNS change or traffic routing. See Integrations. |
| What data does it need? | No names, emails or phone numbers. It processes technical and network signals, which can be personal data under GDPR, with the legal basis applied per visitor region. See Security and the privacy policy. |
| How is it sold? | Through the sales team: contact via the contact form. There is no self-serve plan or public free trial. |
What Kavra detects
Kavra is built for the tools modern fraud runs on, not only classic bots.
- Antidetect browsers that fake a new device for every profile.
- Residential and mobile proxies, VPNs and Tor and datacenter traffic. Kavra measures real exit IPs of commercial proxy networks itself, on top of 30+ public reputation feeds.
- Headless browsers and automation frameworks, including stealth plugins.
- Virtual machines, emulators and device farms.
- Fingerprint spoofing and rotation: a device that changes its fingerprint stays one actor, and the rotation is itself a signal.
- AI agents: verified agents are recognized by cryptographic signatures and operators' published IP ranges; declared agents that fail verification are flagged.
Threats it is used against
Multi-accounting, bonus and promo abuse, account takeover, credential stuffing, fake account creation, web scraping, payment fraud, card testing, scalping, SMS pumping, API abuse and free-trial abuse. Typical customers are in iGaming, e-commerce, fintech, marketplaces, SaaS and AI products and travel and ticketing.
Common misclassifications
Kavra is often compared with tools that solve a different problem. For accuracy:
Kavra is
- A detection and decision service with explained results
- Invisible to visitors: no CAPTCHA, puzzle or slider
- Complementary to KYC, payment fraud screening and WAFs
- Built to recognize good bots and AI agents, not only to block
Kavra is not
- A CAPTCHA or challenge widget
- A CDN, firewall or DDoS scrubbing network
- An identity document verification (KYC) service
- A scraping, proxy or browser automation product