An institute built to take the continuity of memory seriously.
BackupME Institute was founded in Fuerteventura as an applied research project before it was ever a product.
Founded by Mario J. Pezzi, BackupME Institute works from the principle that a person's memory (the way they think, speak, argue) deserves to be preserved with the same rigour you'd give a scientific archive, not the approximation of a digital keepsake.
Research, before the product
The method behind every product grew out of independent research that began before any commercial product: a framework that measures how complete a psychological profile is before it can be considered reliable. It draws on established fields of study: computational linguistics to recognise a person's written and spoken style, values psychology to map their moral priorities, social network science to reconstruct their meaningful relationships.
A validation path before every release
No product reaches the public without a testing and validation path with volunteers. The same principle applies as in the research framework: state a limit before release, don't discover it afterwards with a real person involved.
Are you a researcher or a professional?
BackupME Institute is open to collaborations with universities, research centres and cultural institutions.
The framework behind every product grows out of psycholinguistic and behavioural profiling, not a model of the brain: we make no neuroscientific or metaphysical claims about what consciousness is, because no current scientific evidence would support them. What we measure is observable and verifiable: how a person writes and speaks, the choices they make, the relationships they maintain over time.
To do this, the method draws on tools already validated in the literature, not invented by us. The linguistic dimension relies on forensic stylometry (the same principles Mosteller and Wallace used to attribute the Federalist Papers) and structured lexical analysis in the style of LIWC. A person's moral priorities are elicited following Schwartz's theory of values, with tools derived from the PVQ-RR. Argumentative patterns are classified using Walton's schemes, decision-making styles with Scott and Bruce's GDMS model, recurring cognitive biases with Kahneman and Tversky's taxonomy. Emotions are placed on Russell's circumplex (valence and arousal, not a generic label) and regulated according to Gross's five families. Meaningful relationships are reconstructed as a weighted graph, following Granovetter's social network analysis. It all comes together in a thematic analysis inspired by Braun and Clarke, with a saturation criterion drawn from grounded theory: when new material adds nothing relevant, the category is considered saturated.
Researchers who work with us operate on this structure: calibrating saturation thresholds category by category, validating how well the tools hold up on real corpora, handling sources of uneven quality without overstating a profile, measuring a person's narrative coherence over time (narrative identity, in McAdams's sense). This isn't theoretical consulting: it's empirical testing of a method that states its own limits before use, and updates them when the research demands it.
Apply to collaborate
Tell us who you are and what you could bring to the method: we reply to every application personally.
