Engineering articles and product updates
How ShieldMarc is built: threat detection internals, architecture choices, mission pieces, and public progress updates.
Showing 4 of 55 resources
What We're Building and When You Can Use It
A June 2026 progress update on the ShieldMarc platform: what we were working on, what was due on 1 July 2026, and how to start a free trial.
Our Mission: AI Tools, Open to All
Why we built ShieldMarc with free AI-powered tools at its core, and what we believe about access, protection, and the future of email security.
How ShieldMarc Threat Detection Works
Every DMARC report row is evaluated by a deterministic multi-signal scoring engine to separate genuine threats from noise. AI explains a sender on request. Here is how it works.
Why Modern DMARC Monitoring Costs a Fraction of Legacy Tools
Legacy DMARC platforms carry years of technical debt. We built ShieldMarc from scratch in 2026 with modern tooling and zero bloat.
Why we publish this
Security products ask customers to trust claims they cannot easily verify. Publishing how the system works is the closest thing to a remedy: if we describe the detection logic, the architecture and the trade-offs we chose, a technical reader can judge whether the claims are plausible instead of taking a marketing page at face value.
These articles cover how DMARC aggregate reports are parsed and stored, how threat scoring distinguishes a genuine spoofing attempt from a forwarding artefact, and where the architecture makes deliberate compromises. They also cover mistakes, because a post that only describes decisions that worked is not an honest account of building anything.
Product updates are recorded here as they ship rather than collected into periodic announcements, so the history is a record of what actually changed and when. Readers evaluating ShieldMarc against alternatives are the intended audience as much as existing customers.