MEDJEV is an independent, JEV-inspired research implementation for typed evidence decisions: a clinical record or literature passage, a runtime question and a candidate set supplied at call time, scored into an explicit probability distribution plus calibration scope. Alpha research code, not clinical software.
Organization repository on GitHub
Independent JEV-inspired System One-style typed decision research for clinical evidence, biomedical NLP, calibrated probabilities, and sleep signals
Repo stats from the GitHub API, cached Oct 2026.Homepage
Clinical text · Biomedical literature · Sleep signals · Auditable benchmarks. Research implementation for JEV-inspired evidence reasoning.
Quoted from the project's README. Read the full README.
Watch out
The repository says plainly that it is a research prototype, not the official TypeSafe Jev API, SDK or hosted service, that it does not reproduce proprietary weights, and that it must not be used with identifiable patient data or in patient care. Its own boundaries section adds that evidence grounding is not factuality proof, that calibration is schema-specific and must not be carried to a new candidate set, and that permutation invariance tests the interface rather than clinical generalization. The output is a relationship to supplied evidence, not a diagnosis, prognosis or risk score, and the status badge reads alpha. README ends mid-table on the included packages, so the sleep extension is described only at a high level. Copy reviewed against the repository README on 2026-10-04, not from running it.
Argos1111/jev_localPython
rorshopping/jev-on-a-laptopPython
rlaope/jevalPython
erendikmenn/jev-rag-benchmarkPython
sgoedecke/system-onePython
mithalouni/system-one-openPythonOriginal write-ups on what this entry is for.
Posts from builders shipping with Jev right now.