Elemendar

Elemendar builds natural-language-processing models that read unstructured cyber threat intelligence (CTI) reports, blog posts, and PDFs, then convert the prose into structured, machine-readable STIX 2.x objects mapped to MITRE ATT&CK techniques. The core problem it targets is real and well-documented in the CTI field: most threat intelligence still arrives as human-written narrative, and analysts spend a large share of their time manually re-keying that narrative into the structured formats that SIEM, SOAR, and threat-intel platforms actually consume. Elemendar automates that translation step rather than generating the intelligence itself.

The company was founded in 2017 by Giorgos Georgopoulos and Syra Marshall out of the UK National Cyber Security Centre-backed Cyber Accelerator (Wayra UK), and is based in London. Its flagship product, READ, ingests CTI documents and outputs STIX bundles with ATT&CK tagging, intended to plug into existing threat-intel platforms (MISP, ThreatConnect, and similar) rather than replace them.

Elemendar is a small, early-stage company that has raised a modest amount of outside capital relative to the broader AI-security market, and it operates in a narrow niche (CTI structuring) rather than a broad security category. That focus is also its main limitation: the value of READ is capped by how much of an organization’s workflow still depends on unstructured intelligence versus feeds that already arrive structured.