Applied Cryptographers
Develop advanced cryptographic schemes such as homomorphic and functional encryption, deniable and anamorphic communication, and verifiable ML primitives — translating cutting-edge research into deployable systems.
Research-driven civic cryptography studio
Cryptography for fair, private and democratic digital systems.
Transpara Labs builds cryptographic systems that allow institutions to
analyse data without seeing it, evaluate individuals without identity or
demographic leakage, verify the integrity of machine-learning models, and
communicate safely under surveillance or repression.
Our mission is to ensure that digital infrastructure respects fairness,
privacy and human rights by embedding these values directly into its
cryptographic foundations.
About Transpara Labs
A research-driven cryptography studio for democratic societies.
Transpara Labs is a European research and technology studio developing cryptographic systems that protect individuals, strengthen institutions and support democratic processes. Our work spans privacy-preserving analytics, digitally blind evaluation, verifiable machine learning and advanced communication schemes for high-risk or repressive environments.
We work at the intersection of applied cryptography,
fairness and governance and
human-rights-centred design. From homomorphic and
functional encryption to deniable and anamorphic communication,
we convert advanced theory into practical tools that institutions and
communities can rely on.
How we work
We collaborate with universities, public-sector organisations, NGOs and research consortia to design and evaluate new cryptographic mechanisms. Our approach is pragmatic: prototype quickly, validate with real users and environments, and refine systems until they are robust, usable and deployable.
What we value
We believe that fairness, transparency, privacy and safety are foundational to digital societies. Our systems embed these values cryptographically, enabling decision-making processes and communication channels that remain trustworthy even in sensitive or high-risk contexts. We favour open standards, verifiable protocols and rigorous scientific foundations.
Our Expertise
Disciplines we bring together.
Transpara Labs combines deep cryptographic research with domain expertise in governance, human rights and real-world deployment. We work as a small core team supported by a flexible network of collaborators, assembling project-specific constellations of expertise rather than a fixed, static team.
Privacy-Preserving ML
Combine cryptography and machine learning to enable analytics and inference on encrypted or pseudonymised data. We specialise in fairness-aware approaches that reduce reliance on demographic proxies.
Verifiable & Accountable AI
Design mechanisms that ensure ML predictions come from certified models, remain tamper-evident and comply with governance rules. Includes model attestation, verifiable inference and trustworthy deployment pipelines.
Secure Communication & Resilience
Work with journalists, activists and civil-society organisations to design communication channels that remain safe under surveillance, coercion or repression — informed by real threat models and field constraints.
PQC & Cryptographic Consultancy
Support organisations transitioning to post-quantum cryptography and improving their cryptographic posture. Includes protocol redesign, migration planning and compliance with emerging PQ standards.
Education & Capacity Building
Create cryptography, privacy and security training for companies, universities and schools. We also develop accessible educational tools inspired by the "Alice in CryptoLand" video game.
Products
Operational products built on our cryptographic research.
Beyond research prototypes, Transpara Labs builds production-grade tools for institutions that need strong security with strict privacy and civil-liberties guarantees.
Privacy-first security intelligence
Blind Brother
Blind Brother is a selective video-encryption platform for modern camera environments. It lets operators protect privacy by default while preserving useful evidence and auditability.
- Real-time face, human and vehicle encryption on live streams.
- Role-based selective reveal with audit-ready access logs.
- Public-safety workflows that protect civil liberties by design.
Coercion-resistant secure messaging
The People's Signal
The People's Signal applies anamorphic encryption to messaging, embedding overt and covert content into one ciphertext so users can preserve plausible deniability under coercion.
- Dual-message ciphertexts for overt-safe and covert-sensitive communication.
- Plausible deniability when standard keys are disclosed under pressure.
- Research-to-product path grounded in formal security goals and prototype testing.
Current projects
Systems we are building.
The following projects reflect our ongoing work in applying advanced cryptographic techniques to real institutional and societal challenges. Each prototype explores how formal guarantees can translate into deployable, practical systems.
Urban surveillance, reimagined
Blind Brother for metropolitan camera networks
We are adapting our Blind Brother protocol for deployment in large city camera systems, using attribute-based selective encryption to give different keys and permissions to different parts of a video stream. Faces, licence plates or other regions of interest are encrypted under dedicated policies rather than exposed by default.
The goal is to let operators act on relevant security events while sharply limiting unnecessary exposure of citizens’ identities. Oversight bodies can audit which parts of which streams were actually decrypted, helping answer who, in practice, is allowed to see what in pervasive surveillance settings.
Selective video encryption
Attribute-based access
Privacy-preserving surveillance
Fair decision pipelines
Digitally blind evaluation infrastructure
We are designing end-to-end evaluation systems where committees and machine-learning models never see directly identifying or sensitive demographic information. Applications, proposals or cases are processed so that only the features needed for a decision are revealed, while identity and potential proxies remain hidden.
The project targets grant allocation, hiring, admissions and similar high-stakes selection processes. It aims to provide cryptographic guarantees about which attributes influence outcomes and which do not, enabling institutions to implement fairness-by-design instead of relying solely on ex post audits.
Digitally blind evaluation
Fairness by design
Encrypted decision-making
Secure communication under coercion
Anamorphic messaging for high-risk users
We are building a messaging prototype based on anamorphic encryption, in which each ciphertext can decode to a coercible “cover” message as well as a hidden one. Under pressure, users can reveal keys that only open the benign layer, while the true content remains indistinguishable from random data.
The design is guided by the needs of journalists, activists and others who operate in repressive or high-risk environments. The aim is to make coercion and device inspection less effective in practice, and to offer realistic, usable tools that support democratic resilience.
Anamorphic encryption
Deniable communication
Democratic resilience
Publications
Recent papers and accepted work.
Research outputs from our work on privacy-preserving cryptography, secure systems and accountable digital infrastructure.
“It Runs and it Hides” accepted for publication.
Our work on Function Hiding Functional Encryption was accepted for publication at ICISSP 2026. This work creates a provable secure cryptographic construction that, given any message-private functional encryption scheme, yields a function-private one.
“Oops!... They Stole it Again” accepted for publication.
This work systematically reviews various attacks on Split Learning (SL), classifying them based on factors such as the attacker's role, the type of privacy risks, when data leaks occur, and where vulnerabilities exist. We also analyze existing defense methods, including cryptographic methods, data modification approaches, distributed techniques, and hybrid solutions.
“The Sound of Reduction” accepted for publication.
This work presents a novel fuzzing framework that combines complexity reduction with coverage guidance to improve the effectiveness of fuzzers when applied in cryptographic software.
“Blind Brother” accepted for publication.
Our work on privacy-preserving selective video encryption was accepted for publication at CODASPY 2025. The system provides cryptographic guarantees for surveillance in metropolitan environments.
Contact
Collaborate with Transpara Labs.
If you are exploring privacy-preserving analytics, digitally blind evaluation, verifiable machine learning or secure communication in high-risk environments, we are happy to discuss how our work could support your project.
How we usually collaborate
We work with universities, public-sector organisations, NGOs and companies on research proposals, pilots and long-term cryptographic infrastructure. Typical roles include architecture design, prototype development, privacy-preserving analytics, secure communication tooling and dissemination.
Privacy-preserving analytics, digitally blind evaluation, verifiable and accountable ML, secure communication under surveillance, post-quantum cryptography and cryptography education.
You can reach us directly at info@transparalabs.eu.
Tornimäe tn 5, 10145, Tallinn, Estonia
Send us a message
Share a short overview of what you are working on. We will get back to you as soon as we can.