Language is part
of the architecture.
How instructions, context, and interpretation shape the behavior of an AI agent. Our clinical and linguistic background informs the questions we ask.
INDEPENDENT RESEARCH LAB · BRAZIL
We create environments for advanced AI agents and offensive security research. Our work connects linguistics, Lacanian psychoanalysis, and software engineering.
Explore our research01 / OUR RESEARCH
We bring a background in Lacanian psychoanalysis and linguistics to the design of research environments for AI agents and offensive security.
How instructions, context, and interpretation shape the behavior of an AI agent. Our clinical and linguistic background informs the questions we ask.
Environments for advanced agents to investigate security problems. Our work extends into hands-on research, bug bounty submissions, and reproducible findings.
Research environments with tools, context, and continuity. We build the infrastructure that supports extended investigations and lets us revisit the evidence.
02 / IN PRACTICE
Our security research is reflected in our individual Intigriti profiles.
Explore the leaderboardBoth researchers have reached #1 on a program’s quarterly leaderboard.
Intigriti snapshot · . Rankings change over time. Verify Alice’s 90-day position ↗
03 / HOW WE WORK
Our work moves between theory, implementation, and observation. Each experiment gives us a more precise question.
START WITH WHAT HAPPENS
We inspect how agents respond to language, use tools, and move through a task. Recorded sessions and experiment traces let us return to what actually happened.
LANGUAGE CHANGES THE QUESTION
Linguistics and Lacanian psychoanalysis inform how we think about meaning, context, and the position of a speaker. We bring that attention to the design and study of agent instructions.
GIVE THE QUESTION A WORKING FORM
We create research environments with tools, context, and room for an investigation to develop. We study how those conditions shape what agents can discover.
LET THE EVIDENCE ANSWER
We revisit observations, reproduce security findings, and assemble evidence for review. Bug bounty submissions provide an external point of reference for our security work.
04 / THE PEOPLE
Based in Porto Alegre, Brazil. Working across the boundaries of language, human subjectivity, and technology.
Psychologist · Lacanian psychoanalyst
Alice brings her background in psychology, psychoanalysis, and linguistics to the creation of research environments and the study of intelligent systems.
@sippertalice on IntigritiPhD in Social Psychology · Lacanian psychoanalyst
A psychologist and psychoanalyst, Bruna brings a research background in social psychology to the lab’s work on language, agent behavior, and security.
@hans_lupin on Intigriti05 / INSIDE THE LAB
Research shapes our software. Our software opens up new research.
We create environments in which agents can pursue questions, use tools, and sustain long research sessions. Language, context, and continuity are part of the environment’s design.
We preserve and study agent rollouts: the recorded sequence of a session. Successful, partial, and failed runs help us examine how agent behavior changes with instructions and context.
Structured reports, reproducible observations, and retained evidence connect a research question to a result another person can review. Our work spans web applications, APIs, and software infrastructure.
START A CONVERSATION
Research collaborations, technical challenges,
or a conversation across disciplines.
PORTO ALEGRE, BRAZIL · CONNECTED GLOBALLY