AI features in homeopathy software, ranked by usefulness

AI features in homeopathy software range from quietly indispensable to overhyped, and the difference is rarely visible in vendor marketing. Six feature classes ship in 2026 — semantic search, notes-to-rubrics, live transcription, rubric mapping, natural-language materia medica search, and remedy suggestion — and they sort cleanly by what they actually save you in a working consultation. Each one's documented limits sit alongside it.

The feature classes at a glance

1. Semantic rubric search (natural language in, rubrics out)

Semantic search lets you type a symptom the way the patient says it — "headache better lying down" — and surfaces matching rubrics by meaning rather than exact wording. It is the most mature AI feature in the field and the one with the clearest daily payoff: the rubric hunt that used to take minutes of tree-browsing collapses into one query. Similia ships it in the free repertory alongside a keyword mode for exact matching.

• Function: meaning-based rubric retrieval from natural language
• Modes: semantic and keyword, switchable per query
• Payoff: shortens the symptom-to-rubric translation loop
• Skill still required: choosing among candidate rubrics remains the prescriber's call

Well-suited for: every repertorization session — it becomes the default way into the repertory once adopted. Trade-off: paraphrase coverage is imperfect; uncommon symptom phrasings still need keyword search or manual browsing.

2. Case-notes-to-rubrics analysis (your notes, pre-repertorized)

Notes-to-rubrics tools read freeform consultation notes and propose rubrics, flagging Strange, Rare and Peculiar details a tired prescriber can miss. The Similia implementation analyzes a full note or a highlighted selection, suggests rubrics in the chosen repertory, and digitizes handwritten notes by OCR before analysis. It is credit-metered and framed as a suggestion layer: you review and accept each rubric.

• Function: freeform notes in, suggested rubrics out
• Scope: whole note or selected passage; handwriting OCR supported
• Repertory targeting: suggestions land in the repertory you select
• Cost model: Pro feature consuming AI credits
• Review step: each suggestion is accepted or rejected manually

Well-suited for: long chronic-case notes where SRP details hide in paragraph twelve. Trade-off: notes beyond the character limit must be analyzed in selections, and suggestions still require line-by-line review.

3. Live consultation transcription (the consultation writes itself)

Live transcription records a consultation, transcribes it in real time, and extracts symptoms as they are spoken. Similia's Live Audio Mode is in beta: it transcribes, prioritizes SRP symptoms over modalities and locations, and auto-adds matching rubrics at a documented rate of roughly one symptom per two-minute cycle — about thirty rubrics per hour at maximum. The audio itself is not saved; only the transcript and a summary persist.

• Function: real-time transcription plus symptom and rubric extraction
• Modes: in-person (microphone) and online (microphone plus system audio)
• Extraction rate: approximately one symptom per two-minute cycle
• Privacy: audio recording not saved; transcript and summary only
• Status: beta, Pro-only, credit-metered

Well-suited for: keeping your eyes on the patient instead of the keyboard. Trade-off: extraction is deliberately rate-limited, and every auto-added rubric needs post-session review.

4. AI rubric mapping and switching (close is not good enough)

Rubric mapping covers two related behaviours: translating an extracted symptom into the most specific available rubric, and offering ranked alternatives when the first mapping is close but wrong. Similia's extraction avoids overly generic rubrics — broad anxiety or head-pain headings — prefers rubrics carrying fifteen to a hundred remedies, and exposes a switch control that swaps a chosen rubric for a top alternative. Photo-based mapping extends the same idea to visible physical symptoms.

• Function: symptom phrase to specific rubric, with ranked alternatives
• Generic-rubric avoidance: prefers rubrics with 15–100 remedies
• Switching: one-tap replacement from alternative suggestions
• Photo route: image of a visible symptom to candidate rubrics
• Review step: practitioner confirms every mapping

Well-suited for: tightening repertorizations where a near-miss rubric would distort the analysis. Trade-off: mapping quality depends on the repertory selected, and photo input is for rubric mapping, not diagnosis.

5. Natural-language materia medica search (reading at query speed)

Semantic search applied to materia medica lets you query symptoms across multiple sources at once and jump to highlighted passages, rather than reading serially author by author. Similia supports semantic mode across classic authors and owned premium sources — Murphy, Pitt, Mangialavori, Meditative, Griffith, Scholten and Vermeulen where licensed — with results ranked by match strength.

• Function: multi-symptom semantic queries across materia medica sources
• Sources: classic authors free; premium sources when owned
• Output: ranked remedies with full-text highlight navigation
• Modes: also words-in-sequence, same-sentence, same-paragraph
• Payoff: confirmation reading drops from hours to minutes

Well-suited for: confirming a differential against the source literature before prescribing. Trade-off: coverage follows your licences — unowned premium sources are invisible to the search.

6. AI remedy suggestion (useful drafts, prescriber's judgement decides)

Remedy-suggestion features draft a differential from entered symptoms — Vithoulkas Compass's expert system is the longest-standing example, and newer AI implementations extend the idea. This class ranks last because it compresses the most clinical judgement into the least transparent step. Use the output as a draft to interrogate against materia medica and the case, not as a verdict.

• Function: symptom set in, scored remedy differential out
• Implementations: expert-system scoring; newer model-based suggestion layers
• Use: drafts for review, not prescriptions
• Risk: over-trust in an opaque ranking

Well-suited for: drafting a differential to interrogate, especially in teaching settings. Trade-off: the ranking logic is opaque relative to a hand repertorization, and the prescription is yours, not the model's.

FAQ

Which AI features are free to use?

Semantic rubric search is the main free-tier AI feature; Similia includes it in its Free plan. The heavier features — notes-to-rubrics, live transcription, photo analysis — are Pro features consuming metered AI credits across the implementations available. Expect the free boundary to sit at search and the paid boundary at generation.

Is patient data used to train AI models?

Not in the implementation with the most public documentation: Similia states that submitted patient data is not used to train AI models, that it has Business Associate Agreements with its AI providers, and that covered AI data is processed with zero retention. Verify the equivalent statements for any other vendor before adopting their AI features.

Do these features replace repertorization skill?

No. Every implementation includes a mandatory review step where you accept or reject suggestions. The features compress the mechanical work — searching, transcribing, mapping — while symptom evaluation and remedy choice stay with the prescriber.

How are AI features priced?

Two models dominate: subscription gating (the feature exists only in Pro tiers) and credit metering (each analysis consumes credits, with packs sold separately). Similia uses both — Pro subscriptions include 100 AI credits per renewal, with packs from $14.99 USD for 200 credits. Budget for credits if your caseload leans on notes or audio analysis.

Verdict

Semantic search is the feature worth adopting first — mature, free in at least one platform, and paying off in every session. Generation-class features earn their credits only under review discipline: useful drafts, your prescription.

References

Similia (2026) Pricing and Subscription FAQ, Searching for Rubrics (Semantic Search), Using AI: Case Notes to Rubrics, Using AI: Live Audio Mode (Beta), Using AI: Photos to Rubrics, Studying the Materia Medica, Is my patient data secure?, similia.crisp.help, fetched 2026-05.

Vithoulkas Compass (2026) product pages, vithoulkascompass.com, fetched 2026-05.