The only email gateway that clusters active-active without a master.

Self-hosted. Open-source stack. GDPR by architecture. NetCell MailGuard places itself in front of your existing mail server via MX record — spam, phishing and malware are filtered locally on your servers. No cloud, no US third parties, no data export.

curl -fsSL https://get.netcell-mailguard.de | sudo bash

Debian 12/13 · amd64 + arm64 · Single-node or cluster

NetCell MailGuard admin dashboard — mail traffic, threat radar, Bayes & neural classifier statistics
0

Cloud dependencies — everything on your servers

Cluster nodes — no quorum, no master/replica

14

Detection layers — from DNSBL to attachment sandbox

GDPR

by architecture — made in Germany, no data export

GDPR by architecture — no compliance acrobatics required.

Cloud providers like Mimecast, Proofpoint and Hornetsecurity route every email through third-party infrastructure. For law firms, tax advisors, hospitals, public authorities and anyone under strict GDPR/BDSG supervision, that means a compliance nightmare with data-processing agreements, third-country transfers and standard contractual clauses. NetCell MailGuard runs on your server, in your data centre — the mail never leaves your infrastructure.

No data-processing agreement You are the processor — nothing goes to third parties. GDPR Art. 28 DPA is not required at all.
No third-country transfer All components stay in the EU/EEA — no Schrems II risk, no standard contractual clauses needed.
Complete audit log Every admin action, every quarantine release, every policy change with timestamp and actor — mandatory evidence for Art. 5 (1) f.
Vendor based in Germany NetCell IT, Leverkusen — German support, German contracts, no US Cloud Act access.

GDPR argument in detail →

Eight features that set us apart.

Every vendor offers SPF/DKIM/DMARC. These eight points are our real differentiators.

Active-active without a master

Every cluster node is an equal peer — no failover, no promote, no quorum. Configuration and state are encrypted and synchronised across all nodes. Scales horizontally without limit.

GDPR by architecture

Self-hosted, no data export, no cloud provider between sender and recipient. No DPA, no third-country clauses. Law firms, tax advisors, public authorities and hospitals are the target audience.

Local attachment sandbox

Suspicious Office, PDF and archive attachments are detonated in an isolated sandbox on your server — not shipped to a cloud sandbox vendor. Operator-managed detection rules.

DACH-phishing detection

Patterns for German-speaking phishing waves: account-locked, tax refund, GEZ, Klarna, DHL parcel, Apple ID. English-trained cloud models often miss these.

End-user quarantine portal

Recipients receive a daily digest with single-click release. No admin involvement for false positives, no „can you release this mail" tickets. Token-based, secure.

White-label branding

Your logo, your product name, your digest subjects as a global cluster identity. Resellers sell MailGuard under their own brand — end customers don't see „NetCell MailGuard".

API-first with OpenAPI

Every UI function is also available via REST API. API keys with scope and rate limit, OpenAPI spec as a versioned contract, webhook hooks for quarantine and reputation events.

Per-domain policies

Quarantine threshold, detection policy, DKIM keys, DMARC reporting, mail filter lists, header rewrites — all configurable per domain. One operator team manages all domains from one UI.

All features in detail →

Every node is master.

No failover drama on hardware failure. No promote script. No split-brain prevention. Every node processes incoming mail, every node replicates configuration changes to all others, every node can join or leave at any time.

   ┌──────────────┐    encrypted    ┌──────────────┐
   │  MailGuard   │ ◀──────────────▶│  MailGuard   │
   │   Node 1     │   sync          │   Node 2     │
   │              │                 │              │
   │  Detection   │                 │  Detection   │
   │  stack       │ ◀──────────────▶│  stack       │
   │  + sandbox   │                 │  + sandbox   │
   └──────┬───────┘                 └──────┬───────┘
          │                                │
          └────── MX round-robin ──────────┘
                       │
                       ▼
              Existing mail server
              (Linux MTA / Exchange / Microsoft 365 / Google Workspace)

Architecture in detail →

30-day trial. After that, EUR 29 per server.

We bill per server, not per mailbox. EUR 29/month — whether 50 or 50,000 mailboxes pass through. Cloud email security vendors (Hornetsecurity, Mimecast, Proofpoint, Microsoft Defender for O365) typically charge EUR 2-5/mailbox: at 1,000 mailboxes that's EUR 2,000-5,000/month.

30-day trial

Full feature set, four weeks to evaluate

EUR 0 for 30 days
  • Full Standard feature set
  • Cluster & multi-domain
  • White-label branding
  • End-user quarantine portal
  • External threat feeds, sandbox
  • No credit card for the trial

Trial activates automatically in the setup wizard — no order process required.

Special requirements like SLA, phone support or on-premise license? Get in touch — we handle these setups individually.

Frequently asked questions

How does MailGuard differ from Hornetsecurity or Mimecast?

MailGuard runs on your own servers — Hornetsecurity, Mimecast and Proofpoint are cloud services. With MailGuard you see every header, every detection decision, and you can audit-log every action. No US third parties, no data hand-off, GDPR by architecture.

How does active-active without a master work?

Every cluster node is an equal peer. Configuration and state are encrypted and synchronised across all nodes. No quorum, no failover, no promote. New nodes join with a single command. Scales horizontally without limit.

Do I need a cloud connection?

No. MailGuard runs on servers or VMs in your own infrastructure. Threat-intel feeds (OpenPhish, URLhaus) are pulled — the mails themselves never leave your infrastructure.

Which detection layers are active?

14 layers: SPF, DKIM, DMARC, ARC, RBL/URIBL, header-anomaly detection, external phishing feeds (OpenPhish, URLhaus, PhishTank), URL-shortener resolver, DACH-specific phishing keywords, suspicious-TLD self-learning, spam scanner with ML classifier (statistical + neural + reputation), virus scanner and sandbox detonation for unknown attachments — all local per node.

Can I resell MailGuard?

Yes, with two caveats: white-label branding (logo, product name, primary colour) is one global identity per cluster — i.e. one brand outward. Per-domain configurable: digest from-address, DMARC rua-address, detection policy. If you need different brands per end customer, you either run one cluster per customer or you live with one unified reseller brand. MailGuard is single-tenant — one operator team manages all domains centrally.

How does the licensing model work?

30-day trial with full Standard feature set, then Standard EUR 29/server/month. No freemium — on day 31 the trial license expires and mail processing pauses. Configuration is preserved for 14 days in case you decide late. Mailbox count is never a limit. Custom licenses with SLA, phone support or on-premise perpetual on request.

Ready for your own server instead of cloud?

Install NetCell MailGuard on your own server. 30 days free, then EUR 29 per server — no credit card required for the trial.

See pricing