SciTransfer
Organization

CYPHEME

French AI SME developing smartphone-based counterfeit detection for packaging, labels, and consumer goods authentication.

Technology SMEsecurityFRSME
H2020 projects
2
As coordinator
2
Total EC funding
€1.5M
Unique partners
0
What they do

Their core work

CYPHEME is a Paris-based deep-tech SME that builds AI-powered counterfeit detection systems, enabling brands and consumers to verify product authenticity using nothing more than a smartphone camera. Their core technology combines computer vision and neural networks trained to recognize security features — including specialized varnishes and coatings — printed or applied to packaging and labels. The company progressed from a feasibility prototype to a commercially deployable platform through the EU SME Instrument, suggesting a market-ready or near-market product. Their solution targets brand owners, packaging companies, customs authorities, and consumer goods manufacturers who need scalable, hardware-free authentication at scale.

Core expertise

What they specialise in

AI-based counterfeit detectionprimary
2 projects

Both Microguard projects (2018 and 2020) are centered on neural network and AI systems for detecting counterfeit products.

Mobile authentication (smartphone-based verification)primary
1 project

The 2020 Microguard Phase 2 project explicitly targets detection via a cell phone camera, indicating a consumer-facing or field-deployable mobile application.

Packaging and label security featuressecondary
1 project

Keywords from the 2020 project include varnish, packaging, and label, pointing to physical security markers on product surfaces that the AI is trained to authenticate.

Brand protection and consumer goods authenticationsecondary
1 project

Brand protection and consumer good protection are explicit keywords in the 2020 Microguard project, positioning their technology for use by manufacturers and retailers.

Border security and customs inspectionemerging
1 project

Border security is listed as a keyword in the 2020 project, suggesting applicability to customs enforcement and import/export fraud prevention.

Evolution & trajectory

How they've shifted over time

Early focus
Neural network counterfeit detection
Recent focus
Mobile brand protection and packaging authentication

CYPHEME's H2020 trajectory follows a textbook SME instrument scale-up arc: a 2018 Phase 1 feasibility study established the concept of a neural network counterfeit detector with no further detail on application domains, followed by a substantially funded Phase 2 project in 2020 that crystallized the commercial focus onto smartphone-based detection, packaging authentication, and border security. The appearance of covid-19 as a keyword in the later project suggests the pandemic context either accelerated demand (counterfeit PPE, vaccines) or broadened the solution's reach into health product authentication. The trajectory moves from pure AI research proof-of-concept toward a sector-specific, deployable product with clear commercial verticals.

CYPHEME is heading toward commercialization of a smartphone-accessible authentication platform, with growing relevance in regulated sectors (health products, customs) where counterfeit detection pressure has intensified post-2020.

Collaboration profile

How they like to work

Role: consortium_leaderReach: regional

CYPHEME has operated exclusively as a solo project coordinator under the SME Instrument, which is designed for single-company innovation development — so the absence of consortium partners reflects the funding scheme, not a deliberate isolation strategy. This means they are self-directed, likely founder-led, and accustomed to driving technical and commercial decisions independently. A consortium partner engaging CYPHEME should expect a technology provider that brings a finished or near-finished product rather than a co-development partner seeking shared R&D.

CYPHEME has no recorded H2020 consortium partners, as both projects were executed under the SME Instrument's single-beneficiary model. Their network within the EU research ecosystem is therefore minimal on paper, though their commercial partnerships and customer relationships — not captured in CORDIS — are likely where their real connections exist.

Why partner with them

What sets them apart

CYPHEME occupies a specific niche at the intersection of computer vision AI and physical product authentication — a space where most players either provide hardware (scanners, readers) or purely digital solutions. Their differentiator is removing the need for dedicated hardware: if their AI can authenticate a product through a standard smartphone camera reading a printed varnish or label, the cost of deployment drops dramatically for brand owners. For consortium builders, they represent a rare combination of deep-tech AI capability in an SME that has already passed EU peer review twice and scaled from concept to commercial deployment within five years.

Notable projects

Highlights from their portfolio

  • Microguard (Phase 2)
    With €1.4M in EU funding, this is CYPHEME's flagship project and one of the highest SME Phase 2 grants in the anti-counterfeiting space, covering full commercial deployment of a smartphone-based AI authentication platform.
  • Microguard (Phase 1)
    The 2018 feasibility project demonstrates a successful Phase 1 → Phase 2 progression — a strong signal of technical credibility, as only a minority of SME Phase 1 recipients secure Phase 2 funding.
Cross-sector capabilities
Consumer goods and retail (brand protection at product level)Health and pharmaceuticals (counterfeit medicine and PPE detection)Customs and trade compliance (border security applications)Packaging and print industry (security feature integration)
Analysis note: Profile is based on two projects with the same name and a thin keyword set. The core technology and commercial trajectory are reasonably clear from project titles and the SME Phase 1→2 progression, but no consortium partner data exists and project deliverables are unavailable, limiting depth. The covid-19 keyword connection is inferred, not documented. Treat sector applications beyond core anti-counterfeiting as plausible extensions, not confirmed capabilities.