Repertorization software, explained
Repertorization software automates one specific thing: the arithmetic that turns a set of selected rubrics into a ranked remedy table. Everything upstream of that arithmetic — case-taking, symptom selection, rubric choice — and everything downstream — materia-medica confirmation, the prescription decision — remains the practitioner's work, exactly as it was when Kentians ruled analysis sheets by hand.
What the machine actually computes
Every repertorization engine, from 1980s RADAR to current web platforms, implements the same small family of combination strategies the tradition developed on paper.
| Strategy | What is computed | Paper ancestor |
|---|---|---|
| Totality / coverage | Count of selected rubrics containing each remedy | Kentian ruled sheet, tick marks |
| Sum of grades | Remedy grades (1–4 or 1–5) summed across rubrics | The same sheet, weighted ticks |
| Weighted analysis | Practitioner-assigned rubric weights multiply the grades | Underlining the case's leading symptoms |
| Elimination | Only remedies present in designated essential rubrics survive | Boenninghausen-style anchor-symptom filtering |
The software is faster, not different. A remedy table is a function of the rubrics fed in: poor rubric selection produces a confidently wrong ranking at machine speed.
The features that actually differentiate products
Once the four strategies are table stakes, five capabilities separate current repertorization software in practice.
Rubric retrieval. Finding the rubric is the slowest manual step, so search quality dominates daily experience. Keyword search is universal; semantic search — retrieval by meaning, in the patient's own phrasing — is the 2020s differentiator, and you can run it against the classical books in the free online repertory alongside keyword mode.
Per-rubric weighting and grading. The interface should make intensity grading (typically 0–4) a one-gesture act, since weights encode the case hierarchy. A usable engine re-ranks the remedy table live as grades change, so the practitioner sees immediately how a single weight shifts the analysis.
Filtering. Mature engines filter the remedy table by kingdom and family, by minimum or maximum grade, and — in source-coded editions like the Complete, Saine, and Suggesta repertories — by author and source. Filters are analysis instruments, not decoration: a family filter operationalises a kingdom hypothesis without rebuilding the sheet.
Case integration. An analysis that lives outside the patient record is half a tool. The standard to expect: analyses attached to cases, notes and prescriptions in the same file, rubric sheets copyable between follow-ups, and export to standard document formats.
AI assistance. The newest layer proposes rubrics from typed notes, photographs, or live consultation audio. These are suggestion engines with real failure modes — a proposed rubric must still be verified against the repertory and against the case before it enters the analysis.
What repertorization software cannot do
Three boundaries hold regardless of vendor.
It cannot select symptoms. Hahnemann's §153 instruction — weight the striking, exceptional, peculiar features — is a judgement about the living case that no ranking function sees. Software ranks what it is given.
It cannot confirm a prescription. The repertory fragments the case into index entries; the materia medica reassembles candidate remedies into portraits for comparison against the whole patient. Kent's instruction that the repertory narrows and the materia medica decides survives every generation of tooling.
It cannot recover a badly taken case. Speed of computation hides the absence of observation; it does not replace it.
Cost shape in 2026
The arithmetic itself is now effectively free. Public-domain repertories with full analysis mechanics are available at no cost — Boericke, Kent and Boenninghausen with semantic and keyword search and a working case file ship on free tiers, and open-source OOReP covers a baseline. Payment buys premium editions and workflow depth: contemporary Pro plans price in the $20–$30 USD per month range depending on whether Murphy, the Complete, or Saine editions are included; legacy desktop suites price their catalogues per licence. For most buyers the deciding cost question is the repertory licence, not the engine.
Verdict
Repertorization software is mature, commodity arithmetic wrapped around a repertory licence. Choose the books and the workflow, verify the export path, and treat every AI rubric suggestion as a draft to be checked against the repertory and the case.
References
Kent, J. T. (1900) Lectures on Homoeopathic Philosophy, Chicago: Ehrhart & Karl; Kent, J. T. (1897) Repertory of the Homoeopathic Materia Medica, Lancaster: Examiner Printing House.
von Boenninghausen, C. (1846) Therapeutic Pocket Book for Homoeopathic Physicians, English translation by C. J. Hempel, New York: Radde.
Hahnemann, S. (1842) Organon of the Medical Art, sixth edition, §153; translation by W. B. O'Reilly (1996), Redmond: Birdcage Books.
van Zandvoort, R. (ed.) (1994 onward) Complete Repertory, IRHIS, source-coded editions.
Saine, A. (2025) Saine Repertory, Materia Medica Pura Project.
Mangialavori, M. (2024) Suggesta 2.1.
Boericke, W. (1927) Pocket Manual of Homoeopathic Materia Medica with Repertory, ninth edition, Philadelphia: Boericke & Tafel.
OOReP (2026) https://www.oorep.com/, fetched 2026-05-17.
Verdict
Ready to act on this?