Clinilattice 2.0 is a ground-up rebuild around real parsers. - HL7 v2, parsed properly: separators read from MSH-1/2 (never assumed), repetitions, components, subcomponents, and the standard escape sequences — with structural findings named per field (missing control ID, bad timestamps, ADT without PID, and more) - FHIR JSON inspected as a resource or a Bundle: a census by resource type, every reference collected, and — inside a Bundle — every reference resolved, so a pointer to a missing entry is a named defect - De-identify before you share: an explicit rule list blanks patient-identifying fields (PID names, identifiers, DOB, address, phones, SSN, kin, insurance) and reports exactly what was removed; untouched fields survive byte-for-byte - Bundled synthetic samples: an ADT^A01 admission, an ORU^R01 result, and a FHIR Bundle with one deliberately broken reference If 1.x disappointed you, 2.0 is the workbench the name promised — test it with your own messages.