Homeo AI vs CuraSim: Traceable AI Repertory or Review Copilot?

Quick answer

Choose CuraSim when a practitioner wants a newly indexed workspace that combines thirteen repertories, Materia Medica, rubric-weighted AI analysis, patient records, diagnostic tools, and human, veterinary, and agro libraries. Choose Homeo AI when the main job is narrower: preserve a completed human case narrative, inspect proposed symptom and rubric structure, open supporting passages, correct the differential, and retain the accepted reasoning for follow-up.

CuraSim's public pages were reviewed on 30 August 2026. The feature, privacy, and terms routes returned HTTP 200 on curasim.vercel.app, and the legal pages are dated 9 July 2026. The site repeatedly presents curasim.app in interface examples and contact addresses, but curasim.app and www.curasim.app did not resolve in DNS during this review. That mismatch does not prove the public Vercel product is illegitimate or unusable. It does mean a buyer should not infer a settled commercial identity, canonical domain, support path, or contracting entity from polished launch copy.

What CuraSim currently presents

The repertory page names thirteen sources and says their original grading systems remain separate. It describes plain-language rubric discovery, direct movement from a rubric to a saved case, and coverage across human, veterinary, and agro practice. The AI page says a narrative can be parsed into rubrics, with Mind and Generals weighted more heavily, candidate remedies repertorised, and leading results checked against Boericke keynotes.

CuraSim displays High, Medium, and Low confidence bands plus percentages. The page carefully says those values represent agreement within the available rubrics and keynotes, not the probability of cure. That is a genuine strength in its public explanation. The practitioner still needs to test how the number is calculated, which source editions are used, how missing or contradictory rubrics affect it, and whether the interface exposes every transformation needed to reproduce the result.

The root page lists a free Student tier with a classical library and up to ten saved cases. It displays a Practitioner tier at INR 200 per month with AI analysis, unlimited patients and cases, diagnostics, and more than twelve languages. The Clinic tier is custom per seat. The site also states that prices exclude 18 percent GST and advertises a three-month vendor trial. These are dated vendor claims. Confirm whether the public amount is a launch price, which currency and tax rules apply to the actual account, what the trial converts into, and whether every named module is live.

Why CuraSim is not just another clinic-suite query

Many clinic products already covered by Homeo AI focus on appointments, billing, inventory, prescriptions, and staff operations. CuraSim's distinctive claim is different. It connects a named multi-repertory source set to an inspectable remedy shortlist and explicitly explains the limits of its confidence value. It also extends the same workspace into veterinary and agro repertories, which makes source boundaries and record separation part of the buying decision.

That breadth can be a strength for a practitioner or institution that actually needs several case types. It can also create governance risk if human patient records, animal cases, crop cases, diagnostic uploads, and community content share roles, retention rules, or exports. A buyer should test each data class separately instead of treating “one workspace” as a universal advantage.

Homeo AI should not be credited with CuraSim's exact thirteen-repertory collection, veterinary or agro scope, diagnostic modules, community, confidence bands, or INR 200 launch offer. CuraSim should not be assumed to reproduce Homeo AI's exact correction history merely because its marketing says every suggestion is traceable.

Public privacy and contract evidence

CuraSim's privacy policy says patient data remains under the practitioner's controller responsibility and CuraSim acts as a processor. It says clinical text and uploads may be sent to an AI provider to generate analysis, reports, mentoring, or moderation. The policy describes encryption in transit, password hashing, access-controlled storage, deletion requests, international transfers, and service-provider categories.

Several procurement questions remain. The public policy does not name the AI, hosting, database, storage, messaging, or email providers. Its training language says third-party model training is not permitted except as needed to provide the service or with consent. A clinic should ask what “as needed” permits, whether providers retain prompts, how opt-out works, which regions receive data, and how backups and support logs are deleted.

The terms say the contract is governed by the laws of the jurisdiction in which CuraSim operates, but the reviewed text does not identify that jurisdiction or the legal entity operating the service. The contact uses legal@curasim.app while that domain did not resolve during this review. Require the vendor's legal name, address, jurisdiction, tax identity, data-processing agreement, support route, and executed terms before using identifiable patient data or paying for a clinic plan.

These gaps should be reported precisely. A missing domain or unnamed jurisdiction does not prove bad security. It prevents the public documents from establishing a complete purchase contract.

How Homeo AI approaches the same job

Homeo AI is designed around a focused human case-review chain. A practitioner starts with a completed narrative, verifies how expressions became structured symptoms, inspects and corrects proposed rubric mappings, opens source passages, compares remedy differences, and preserves the accepted rationale for the next visit.

CuraSim may fit better when a practitioner wants the repertory, records, diagnostics, and broader libraries to live in one environment. Homeo AI may fit better when the bottleneck is the quality and reconstructability of the narrative-to-rubric review itself. Neither feature breadth nor a narrower design is automatically superior. The better product removes the clinic's measured bottleneck without obscuring corrections or governance.

A reproducible practitioner test

Prepare eight completed, de-identified cases whose reasoning is already documented. Use four human cases, one follow-up, one veterinary example only if the evaluator is qualified for that work, one agro example only if it is genuinely in scope, and one deliberately incomplete case that should produce uncertainty. Keep an external answer packet with the original wording, accepted rubrics, source editions, remedy differential, rejected interpretations, and follow-up changes.

In CuraSim, record the original narrative, extracted rubrics, source repertory and edition, grade display, weighting, shortlist, confidence band, percentage, keynote matches, corrections, export, and time. Change one modality and one negation. Verify that the ranking moves for an understandable reason and that the old result remains available. Check that a percentage is not presented elsewhere as a clinical probability.

In Homeo AI, enter the same human cases. Record every proposed symptom and rubric accepted, edited, or rejected, each passage opened, how the differential changes after correction, and whether a second practitioner can reconstruct the accepted review. Reopen the follow-up and confirm earlier reasoning was not overwritten.

For governance, create only synthetic test accounts and records. Test role separation, export, deletion, session expiry, password recovery, failed AI requests, quota or trial transitions, and removal of an uploaded file. Confirm where every data class appears after export and whether deletion also covers attachments, generated reports, logs, and backups.

Score the products on narrative fidelity, source edition, correction time, confidence interpretation, repeatability, export completeness, role control, deletion, and total cost. Do not score a match to the historical remedy as proof of accuracy. A fluent explanation and a familiar repertory result can still come from an incomplete case.

Known limits and clinical responsibility

Thirteen named repertories do not prove that every source is complete, licensed, correctly graded, or suitable for a clinic's method. A visible percentage is not clinical certainty. An official privacy page does not replace an executed data agreement when the entity, jurisdiction, and subprocessors remain unclear. A polished Vercel deployment does not establish a stable production domain.

Homeo AI can also misunderstand a narrative, suggest an unsuitable rubric, omit relevant evidence, or produce a plausible differential that needs correction. A visible source trail improves reviewability; it does not make a conclusion correct.

Neither product replaces diagnosis, referral, an emergency-care decision, practitioner judgment, prescription, potency, dosage, or follow-up responsibility. Urgent, severe, or rapidly worsening symptoms require appropriate medical assessment. The practitioner remains accountable for the original record, every accepted interpretation and source, and every clinical action.

Conditional verdict

Choose CuraSim if the practitioner needs its named repertory breadth and wider workspace, the controlled test confirms that rubrics, weights, keynotes, and confidence changes are genuinely traceable, and the vendor resolves its public domain, contracting entity, jurisdiction, subprocessor, retention, and checkout questions in writing.

Choose Homeo AI if the priority is a focused, reconstructable path from the human narrative through corrected symptom and rubric structure to inspected passages, remedy differentiation, and follow-up. If both are tested, compare the same known cases and measure correction burden, source clarity, data control, and continuity instead of choosing from module count or launch price.

Official sources checked

• CuraSim official product and pricing page
• CuraSim official repertory feature page
• CuraSim official AI case-analysis page
• CuraSim official privacy policy
• CuraSim official terms of service

Product facts reviewed 2026-08-30. Confirm current features, prices, source libraries, and terms with the vendor.