Skip to main content

Security report for toloka.ai

Toloka AI B.V · Technology · Netherlands

Low risk. 3 critical or high findings in GDPR outside view and SSL/TLS.

Security score

85/100

Risk levelLow

0–100, higher is better. How it is calculated

Toloka AI B.V operates the website toloka.ai, providing AI training data and data solutions that integrate human expertise with automation to accelerate AI development. The company is positioned as a technology provider focused on smarter AI agents and models, targeting AI developers and enterprises requiring high-quality training data. The website is professionally designed using the Framer platform and hosted on Microsoft Azure DNS infrastructure, incorporating multiple marketing and analytics technologies such as Google Tag Manager, Facebook Pixel, and HubSpot Analytics. Security posture is adequate with HTTPS enabled and domain registration protections in place, but lacks advanced security headers and DNSSEC. Privacy compliance is weak due to absence of explicit privacy and cookie policies. No direct contact information or incident response details are found, limiting transparency. Overall, the website demonstrates a solid business presence with room for improvement in privacy and security disclosures.

Scanned

NIS2 and GDPR posture

GDPR outside view
80/ 100 · Low
Checks completed
7of 7

NIS2 Art. 21(2)

  1. (a) Risk analysis and security policies: Not visible from outside
  2. (b) Incident handling: One indicator
  3. (c) Business continuity, backups and crisis management: Not visible from outside
  4. (d) Supply chain security: Not visible from outside
  5. (e) Secure acquisition, development and maintenance, including vulnerability handling: Partly checked
  6. (f) Assessing the effectiveness of security measures: Not visible from outside
  7. (g) Cyber hygiene and security training: Not visible from outside
  8. (h) Cryptography and encryption: Partly checked
  9. (i) HR security, access control and asset management: One indicator
  10. (j) Multi-factor authentication and secured communications: Not visible from outside

This scan rated NIS2 with an earlier keyword method that is no longer shown; rescan for the per-article view. Risk management, continuity, supply chain, training, MFA, management oversight (Art. 20) and incident reporting (Art. 23) need internal evidence, and Guard cannot tell whether NIS2 applies to this organisation.

GDPR is indicative only, based on publicly visible signals: Articles 7, 12, 13 and 32 of Regulation (EU) 2016/679 and Article 5(3) of the ePrivacy Directive. Not a certification or legal assessment.

Priority findings

3 findings rated critical or high. Fix these first.

  1. GDPR outside view

    These are the privacy signals visitors and data protection authorities look for under GDPR.

    • High

      No Cookie Policy found

      GDPR requires clear information about cookie usage

      What to do: Add a dedicated cookie policy explaining cookie usage and purposes

  2. SSL/TLS

    TLS encrypts the connection and proves the site is really yours. Weak or expiring certificates trigger browser warnings and put traffic at risk.

    • High

      Weak Protocols Supported

      Server supports weak protocols: TLSv1.1

      What to do: Disable TLS 1.0, TLS 1.1, and all SSL versions

      • CWE-327
      • OWASP A04:2025
      • ASVS V12.1.1
    • High

      SSL Certificate Expires Soon

      SSL certificate expires in 23 days

      What to do: Renew SSL certificate before expiration

      • NIS2 21(2)(h)
      • IR 2024/2690 9.2(c): Implementing Regulation (EU) 2024/2690, Annex point 9.2(c): Cryptography
      • NIST SC-17: NIST SP 800-53 Rev. 5 SC-17: Public Key Infrastructure Certificates

Need hands-on remediation? Email guard@offseq.com

Keep watching toloka.ai

This scan found 12 issues across 6 checks. Monitoring re-runs those checks daily, emails you when something material changes and records each fix, so you can show progress.

Daily monitoring

Recommended

€25per month

  • TLS certificates and expiry
  • DNS records
  • SPF, DKIM and DMARC
  • Security headers
  • GDPR and NIS2 posture
  • Exposed services
  • Third-party trackers
  • Security score drift

Email alerts on material changes and an exportable JSON history of every change.

Billed monthly to your card. Cancel anytime. The full timeline opens with your unlock code.

Full report

€39.90one-time

  • The complete analysis as a PDF you can cite
  • Email delivery with an unlock code
  • Lifetime access to this report

Prices incl. 21% VAT. Secure checkout by Mollie; sold by SEQ SIA, Latvia.

Results by check

Each check with its score, findings and the evidence it collected.

How this score is calculated

The security score runs 0–100, higher is better. It is the plain average of the security checks that completed in this scan, each counted equally. A check that could not run is left out: it counts neither as a pass nor as a fail.

Counted in this score

SSL/TLS
67
Security headers
80
Email security
90
DNS health
90
Network security
100
GDPR outside view
80
Average
85

Not counted

  • NIS2 outside view: Re-weights evidence from the TLS, email, DNS, network and hygiene checks, so counting it would count that evidence twice. Shown on its own.
  • OffSeq Observations: An AI reading of the site, useful context but not reproducible evidence. Shown on its own.

Findings that cap the score at 49

Each of these is serious on its own, so while any is present the site cannot rate better than high risk, however good the other checks are.

  • The certificate is not trusted by browsers
  • The website is not served over HTTPS
  • A database, remote desktop, file-sharing or Telnet service is reachable from the internet
  • The page loads a script from a compromised source
  • The domain registration has expired
  • SPF authorises every server to send mail as the domain

Partial scans

When the site blocks Guard or most checks cannot run, the score covers only the checks that ran (usually DNS and email), is capped at 75 and is shown as partial, without a risk level.

Risk levels

  • Critical0–25
  • High26–49
  • Medium50–75
  • Low76–100

Breach exposure

Leaked credentials and lookalike domains from OffSeq Breach. The counts are free; the detailed lists are paid add-ons.

Dark web exposure

Add-on

Leaked records that mention toloka.ai, from OffSeq Breach.

Total records
283
Employees
0
Customers
283
Third parties
0
Weak passwords
13%

5 subdomains affected

See the 10 newest leaked credentials

A one-off add-on to this report that lists the most recent identifier and password pairs OffSeq Breach holds for toloka.ai.

Select it in the checkout above.

Breach keeps ingesting new leaks, so these numbers can grow between scans.

Technical appendix

Business context, site observations and technical notes gathered during the scan, and the standards behind each reference.

Business context

Business profile

Toloka AI B.V is a medium-sized technology company founded in 2020 in the Netherlands, specializing in AI training data solutions. Their business model revolves around providing data services that combine human and automated processes to enhance AI development. The company targets AI developers, data scientists, and enterprises in the AI sector. The website reflects a consistent brand and professional content but lacks visible trust indicators such as certifications or testimonials. The marketing ecosystem includes major ad networks and tracking services, indicating active customer acquisition efforts. The transparent WHOIS data supports legitimacy and domain age aligns with company history. No related or partner domains were identified from the provided data.

Security posture

The website employs HTTPS and domain registration protections, which are positive security indicators. However, the absence of DNSSEC and security headers such as Content Security Policy, HSTS, and X-Frame-Options reduces the overall security posture. No vulnerability disclosures or incident response contacts are published, which may hinder timely security communication. The use of multiple third-party tracking scripts increases the attack surface and privacy risks. No exposed sensitive data or known vulnerabilities were detected in the content. The security score is moderate, with recommendations to enhance DNS security, implement security headers, and publish security policies.

Recommendations

  1. Enable DNSSEC on the domain to protect against DNS spoofing and improve DNS security.
  2. Implement standard security headers including Content Security Policy (CSP), HTTP Strict Transport Security (HSTS), X-Frame-Options, and X-Content-Type-Options.
  3. Publish comprehensive privacy and cookie policies with clear GDPR compliance and consent mechanisms.
  4. Add explicit contact information including security incident response contacts and consider publishing a security.txt file.
  5. Review and minimize third-party tracking scripts to reduce privacy risks and improve page performance.
Site observations

Automated reading of the public site. Ratings are indicative.

Business

Company
Toloka AI B.V
Description
From agentic skills to coding and AI safety — we build data solutions integrating human expertise and state-of-the-art automation to accelerate AI development.
Key services
AI training data, Data solutions integrating human expertise and automation
Content quality
good
Branding
consistent

Technology

Stack
Google Tag Manager, Facebook Pixel, Twitter Ads, Reddit Ads Pixel, LinkedIn Insight Tag, Bing Ads, HubSpot Analytics, Framer (CMS/Website Builder)
Frameworks
Framer
Performance
moderate
Mobile
good
Accessibility
basic
SEO
good

Security

Observed score
75/100
Good practices seen
  • HTTPS enabled
  • Domain status includes clientDeleteProhibited, clientTransferProhibited, clientUpdateProhibited

Analytics and tracking

Services
HubSpot Analytics, Google Analytics (via GTM)
Tracking level
moderate
Privacy practices
poor

Advertising

Ad networks
Facebook Ads, Twitter Ads, Reddit Ads, LinkedIn Ads, Bing Ads
Tracking pixels
Facebook Pixel, Twitter Pixel, Reddit Pixel, LinkedIn Insight Tag, Bing UET
Marketing tools
HubSpot Analytics
Transparency
basic

Site quality

Design
good
User experience
good
Content relevance
good
Navigation
good
Professionalism
good
Trustworthiness
moderate

Key observations

  1. Website uses multiple third-party marketing and analytics scripts
  2. No explicit privacy, cookie, or terms of service policies found in the HTML content
  3. No contact emails, phone numbers, or physical addresses explicitly found in the provided content
  4. Website is built with Framer and hosted on Azure DNS infrastructure
  5. Security headers and advanced security policies are not visible in the HTML content
  6. No vulnerability disclosure or security.txt file detected
  7. WHOIS data is transparent and consistent with business claims
  8. Website content is professional and relevant to AI training data services
Technical notes

Notes

The website is built on the Framer platform, leveraging modern web technologies and hosted on Microsoft Azure DNS infrastructure. It integrates multiple marketing and analytics tools such as Google Tag Manager, Facebook Pixel, Twitter Ads, LinkedIn Insight Tag, and HubSpot Analytics. The site is mobile optimized with good SEO meta tags and Open Graph data for social sharing. Performance is moderate, with potential improvements by optimizing third-party scripts. Accessibility features are basic but present. The domain WHOIS data is consistent and transparent, supporting business credibility. No CMS other than Framer was detected. Overall, the technical infrastructure is modern but could benefit from enhanced security configurations and privacy compliance improvements.
Standards and sources

Reference chips under each finding point to the standard, regulation or specification the finding relates to. They are taken from the official datasets below and mapped conservatively: a finding with no defensible link shows none. NIS2 and GDPR references mark where a finding is relevant evidence, not a compliance verdict. Implementing Regulation (EU) 2024/2690 binds digital infrastructure and digital service providers; for other entities it is the most detailed public reading of the Article 21 measures. WCAG, EN 301 549, European Accessibility Act and Web Accessibility Directive references name the requirement an automated accessibility rule relates to; they are not a conformance verdict.

Retrieved 2026-10-25

Reviews of this site

Share your experience with toloka.ai. Every review is verified by email and published after moderation.

Overall rating
Select a rating
4000 characters remaining

We’ll email you to confirm your review and keep your details private.

Community rating

No published reviews yet. Be the first to review toloka.ai.

Feedback on this report

Rate this report, add an optional comment, and tell us if you want a security specialist to follow up.

Overall report quality
Select a rating
2000 characters remaining

We only use your feedback to improve Guard reports. Contact details are never shared.