Open-source infrastructure
for biomedical AI.
Openbtk is the open-source toolkit that ingests, de-identifies, and grounds biomedical data for LLMs — built for healthcare enterprises, developers, and researchers building next-generation GenAI applications.
Openbtk
The open-source core toolkit for biomedical AI.
Openbtk
Pre-AlphaIngest, de-identify, chunk, embed, retrieve, guardrail, and audit biomedical data for LLM applications.
Explore Openbtk →Why Openbtk exists
Openbtk is an open source library for building biomedical AI applications — turning EHR data and clinical text into model-ready, de-identified, auditable inputs, so teams stop re-solving the same infrastructure problems.
No claim without a benchmark
We don't advertise features that aren't built, or numbers that aren't measured. If something is a design target, it's labeled as one.
Standards-first
FHIR, DICOM, OMOP, HL7, LOINC, SNOMED CT, UMLS — we build on top of existing biomedical standards, not around them.
Open source, on purpose
Runs on a laptop or behind a hospital firewall. No forced vendor lock-in for the tools that shouldn't require it.
Documentation ships alongside Openbtk v0.1
Openbtk is pre-alpha today, so there's no reference docs site yet — a marketing page can't substitute for real API docs once there's a real API. Until then, the README is the source of truth.