Kirha Information

KIRHA builds sovereign AI agents for regulated work, starting with life sciences. Our agents search the evidence, write the code and prepare the documents—and leave a record people can verify, audit and sign. Our mission is to accelerate human coordination so that scientific progress improves lives sooner. Agents take on more of the work. People keep the judgement, the authority and the signature. Regulated work runs on people being able to rely on one another. Someone does the work so that someone else can make a decision, and then signs it. We believe AI can take on much of that work. But nobody should have to sign what they cannot verify. So our agents build the audit trail as they go. Every run is logged, every figure is sourced, and every output traces back to the data and the steps that produced it. They run on certified infrastructure (SOC 2 Type I, ISO 27001, HDS, GDPR), and data stays inside compliant environments end to end. Our foundations make this possible: a curated network of 50+ data providers our agents query directly (FDA FAERS, EudraVigilance, PubMed, EUR-Lex, ANSM, HAS and more); KIRHA Search, an auditable routing layer that speaks each provider's native query language, with provenance on every result; and OpenSAS, our open-source interpreter for the SAS language, so our coding agents can run and improve the code they write. Today our agents work in clinical data programming and CDISC conformance, pharmacovigilance and signal detection, regulatory intelligence, and clinical evidence synthesis. Based in Paris - San Francisco.