Simsbuddy is not another generic chatbot wearing a stethoscope. Its public proposition combines three things most competitors offer only one or two of: separate, examination-specific case banks spanning the RCGP SCA, PLAB 2, MRCP PACES, MRCPsych CASC and UKMLA rather than one undifferentiated OSCE library, voice-based AI simulation alongside traditional patient scripts for partner practice, and a marketplace where independent tutors can review recorded AI sessions or provide live tuition, sitting inside an institutional product with private portals and cohort analytics. That combination is genuinely distinctive. Whether it is equally strong across all five examinations, and what its public materials can and cannot actually prove, is this review's job to separate out.
The three components, and what each contributes
Expert-written case banks: the platform's public materials describe more than 100 expert-written cases distributed across its five major examination categories, a meaningful breadth advantage over single-exam competitors and a figure worth treating as provider-reported and dated at the point you check it, since case counts in this category change and the depth within each examination category deserves scrutiny independent of the headline total. Optional voice simulation: purchased through credits, with one credit representing one minute of simulation and credits that do not expire, a genuinely flexible pricing structure whose total cost still needs calculating against your actual usage pattern rather than compared on subscription price alone. And automated examination-specific feedback: analytics tied to the specific marking domains of whichever examination a case belongs to, rather than one generic feedback template applied across every station type.
The five examinations, reviewed separately
Because the exams themselves are not interchangeable, this platform's fit differs meaningfully across them. RCGP SCA: 12 simulated consultations of 12 minutes each, undertaken remotely from a GP surgery, a format where sustained voice conversation and longitudinal management reasoning matter most, territory well suited to a conversational AI patient. PLAB 2: 16 eight-minute scenarios, a tighter format rewarding focused data gathering and prioritisation over exhaustive history taking, where the platform's exam-specific feedback needs to actually reward efficiency rather than checklist completion to be genuinely useful. MRCP PACES: eight encounters across five stations assessing physical examination, recognition of physical signs, communication, clinical judgement and integrated consultation, and this is where a voice-only platform's structural limitation is sharpest, since no conversational system can supply a real murmur, a real abdominal mass or a real neurological deficit to examine, a distinction covered fully in the dedicated PACES-specific analysis below. MRCPsych CASC: 16 stations split between history and risk assessment, examination, and management, seven minutes each, where the platform's coverage needs particular scrutiny for how it handles risk, capacity and emotionally complex management rather than straightforward history taking alone. And UKMLA and CPSA: worth flagging that no single national UKMLA OSCE exists, individual medical schools set and deliver their own CPSAs under GMC quality assurance, so any "UKMLA-ready" claim from this or any platform maps to the national content map rather than to any specific school's actual local station design, a distinction this cluster treats in full elsewhere.
The traditional scripts, and whether they justify the subscription alone
Simsbuddy's conventional patient scripts for partner practice sit alongside its AI simulation as a separate, non-credit resource, and this matters for candidates weighing whether the subscription earns its cost without heavy AI usage: a candidate using the expert-written scripts primarily for peer role-play, reserving AI credits for solo practice or specific weak areas, gets meaningful value from the case-writing quality alone, independent of how much they end up spending on voice minutes.
The analytics, examined critically
Three questions worth asking of any platform's dashboard, this one included: are scores genuinely comparable across attempts, or does variability in AI-generated conversation make attempt-to-attempt comparison less stable than a fixed human-marked mock would be; are weaknesses identified at a level specific enough to act on, by clinical topic, examination domain or interpersonal behaviour, rather than a single aggregate score; and does the platform distinguish a first, unseen attempt from a repeated, increasingly familiar one, since a rising score on a repeated station measures familiarity with that specific station at least as much as genuine skill improvement.
The human-tutor marketplace
The workflow: complete an AI patient station, receive instant automated feedback, then optionally select a session for external review, granting a listed independent tutor access to the recording and transcript for written feedback or a booked live session. Simsbuddy states that listed tutors are independent educators the platform vets, and that recording is disabled by default until the user actively enables it, a genuinely sensible privacy default worth confirming remains in place at the point you sign up. Worth asking before relying heavily on this feature: whether tutors receive standardised marking guidance and calibration training, what quality-monitoring and complaints process exists, and whether a reviewing tutor sees the AI's own score before forming their independent judgement, since visibility of a prior score risks anchoring a human reviewer toward agreeing with the machine rather than assessing the transcript fresh.
Realistic cost, calculated properly
Three usage patterns worth pricing out before subscribing: occasional AI use, a handful of sessions monthly, where the credit model's flexibility genuinely favours low-volume candidates over a flat subscription; three consultations weekly across a revision period, where the minute-based credit cost should be calculated against comparable competitor pricing directly rather than compared to headline subscription figures alone; and daily practice during the final pre-exam month, the heaviest usage pattern, where high-volume candidates specifically should total the realistic monthly credit spend rather than assume the advertised flexibility is automatically the cheapest option at scale.
What Simsbuddy cannot prove from its public information
Named directly, because a platform this evidently well-built deserves precise rather than vague scepticism. Pass-rate improvement: no publicly visible head-to-head educational-outcome study establishes that using Simsbuddy specifically raises examination pass rates. Accuracy of automated marking: no independent validation of how closely the platform's automated feedback tracks a real examiner's judgement has been published. Equivalence to a calibrated examiner: a materially different and stronger claim than producing plausible-sounding feedback, and one the current public evidence does not establish. Transfer to real patients: whether strong performance against the AI translates to strong performance in an unseen human encounter remains untested publicly. And reproduction of physical signs: a structural limitation rather than a temporary gap, since no current voice-only architecture supplies genuine tactile or auscultatory findings, a point the PACES-specific review develops in full.
The editorial verdict
Simsbuddy's most convincing proposition is breadth and flexibility, five distinct examination case banks, a genuinely useful hybrid of AI and human feedback, and a credit structure that adapts to variable usage. Its weakness is structural rather than a simple quality gap: breadth across five very different examinations does not automatically imply equal depth or fidelity in each, and a candidate should verify the specific examination category they need against this review's exam-by-exam breakdown rather than assume the platform's overall polish transfers evenly everywhere.
Frequently asked questions
Is Simsbuddy worth using for PACES specifically?
For communication, consultation structure and clinical-judgement rehearsal, plausibly yes, and for physical examination and sign recognition, no voice-only platform substitutes for bedside teaching and real patient contact, the distinction the dedicated PACES review works through fully.
How should a candidate verify a disputed management answer from a Simsbuddy station?
Against current UK guidance and the exact product SmPC directly, rather than treating the simulation's own mark scheme as a clinical source, the bridge workflow this cluster's iatroX-connection article sets out in detail.
Does the tutor marketplace remove the need for other human feedback?
Not necessarily: it adds a genuinely useful, selective layer of expert review to AI practice, and candidates preparing for high-stakes examinations should still seek full human-observed mock circuits alongside it, particularly for exams with a physical or holistic component this platform cannot fully simulate.