Oscer
Oscer is an AI virtual patient platform from Melbourne for medical, nursing and allied health students preparing for objective structured clinical examinations and US clinical skills exams. A learner interviews a virtual patient in natural language, by text or voice, and the patient keeps a memory of the conversation, so questions can come in any order. At the end the learner gets a score against a detailed rubric for the case, can reveal hints and repeat it, and tracks improvement against peers. Oscer Prime holds hundreds of adult medicine cases, multiplayer cases for two or more players and a Condition Explorer that generates unlimited quizzes across more than 300 conditions, with guided histories, examination checklists, long cases and Anki integration announced.
Universities get curriculum matched virtual patients, custom patients built with open chat that Oscer puts at over 95% accuracy, school wide OSCE style assessment and reporting for tutors and students. The University of Melbourne and the University of Sydney are named partners, and Oscer says students from more than 150 universities in more than 35 countries have used it. Dr. Tom Kelly founded the company, which later built the Heidi documentation product, and Oscer still runs from Fitzroy in Melbourne.
Prime costs US$6.99 for a month, US$34.99 for six months or US$59.99 for a year at promotional rates, against list prices of US$9.99, US$44.99 and US$69.99, and signup is free with no card.
Capability Axes
An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read
The virtual patient is the model at work. It holds context across a conversation, answers open questions the way a patient would and reacts to the order a student asks in, which a fixed quiz cannot do, and that solo chat is the product's signature. Prime also carries tools that keep value without it, from multiplayer cases run between friends to rubrics, the Condition Explorer's quizzes and progress tracking. The AI is the engine of the core simulation module inside a broader study platform for clinical reasoning.
A learner works through a case alone and receives automated scoring and feedback, with hints and the case rubric to judge the answer and tutor reporting afterward. Oscer's technology page claims "Oscer only makes a mistake once and never again. Accuracy is inevitable." A student practicing solo takes the virtual patient's answers and the score at face value before any tutor sees them, and nothing places a check between a wrong answer or a wrong mark and the student's study. No point is defined where a case stops for correction, and no route exists for escalating disputed feedback, so the claim stands in for oversight.
The virtual patient is built to keep context and memory across a conversation, so it answers open questions instead of following a fixed script, and Oscer says the system is more than question mapping or intent search, taking current research into production code. A medical knowledge graph built from clinical reasoning and case data, which every student's use helps grow, sits behind it. Scoring runs against case rubrics. The model family behind the patient, the voice components, which interactions are used to train it and how a change in the model or the scoring is checked before release are never named.
Oscer's privacy policy lets it share data with service providers where needed to run the application, and it offers sign in through outside identity services, which handle access rather than the simulation. The parties behind the virtual patient's chat, the voice mode and the scoring are not named.
Learner conversations and activity are collected in the service, and the about page says every student who uses Oscer helps build its medical knowledge graph, so interaction data feeds Oscer's own models. Whether any outside model provider receives that text, and under what retention or training terms, is left open. Schools receive study data under separate conditions for reporting or optional research.
Institutional partners named are the University of Melbourne and the University of Sydney, and Oscer counts students from more than 150 universities in more than 35 countries. Its technology page claims over 90% accuracy across 200+ conditions and millions of unique inquiries, and the institutions page claims over 95% accuracy for open chat in custom patients. What counts as a correct answer, who judged it, how many interactions were tested and when are never given. Student testimonials describe more confidence and flexible practice. Nothing measures whether Oscer practice raises OSCE scores or improves clinical reasoning.
The privacy policy covers account information, chat logs and in app activity, keeps information while it is needed to provide the service and while the learner holds an account, and allows anonymized or aggregated data to be used for research, development and marketing. Students can opt into research studies with their own sharing conditions, and the about page says student use helps build Oscer's medical knowledge graph.
No fixed retention period applies to a simulation transcript, nothing limits using learner conversations to train the models, and no procedure covers investigating or reporting incorrect simulated clinical content. The data at stake is students' study behavior rather than patients' records.
The platform handles educational accounts, conversations with simulated patients and study progress under a privacy policy that reaches the service, covering what learners enter, chat logs, study responses and activity, and explaining what goes to a research institution or a student's school when they opt in.
No business associate status is stated, and the policy never says plainly that the simulator neither needs nor accepts real patient information, which would be the scope position for an education product. The cases are fictional and the terms frame them as educational, but the posture rests on that framing rather than an explicit statement.
Oscer's terms ask students to protect their credentials and say what to do if an account is compromised, and the privacy policy describes how account and simulation data is used and how the team contacts learners about security issues. Universities see reporting only for their own users. No security certification, independent testing schedule or audit is named for the platform or the student data it holds, and no trust page exists. Sign in through email or outside identity providers describes access, not protection. For universities feeding cohort performance data into the service, the safeguards rest on the company's word.
Education is the product's stated purpose. Students practice history taking, build differential diagnoses and compare their work with a rubric, and educators align virtual patients with a curriculum and review results. The terms limit the content to education and state that the service gives no medical advice, and every case involves a simulated patient rather than a real one. The about page's mission to eliminate missed diagnoses is aimed at training clinicians, not diagnosing patients. No device clearance is claimed, and the product matches the training purpose and setting the company states.
Tutors choose curriculum matched cases and review feedback in institutional reports, which gives a school control over what students practice. Oscer claims over 90% accuracy for its technology and over 95% for open chat in custom institutional patients.
No evaluation compares the quality of the virtual patient's answers or the fairness of its scores across accents, first languages, learner levels or disciplines, and nothing describes how a new virtual patient is tested for misleading or biased answers before students see it, or how marking quality is watched when a school changes its case set.
Learners see the rubric and score after each case, and tutors use institutional reports to review student performance, so a learner or educator can check how a case was marked against its criteria. The terms disclaim the accuracy, completeness and timeliness of the content and any particular results from use, and support covers login and account questions.
No accuracy commitment stands behind the virtual patient's answers, no process exists for appealing a simulation score, and no correction route is described for educational content found to be wrong, a gap that matters for a tool students use to prepare for examinations.
Integration is not part of the product. Cases run in a browser, participation and results are tracked for students and tutors, curriculum matched cases are configured with the institution, and students sign in by email or an outside identity service. No electronic health record connection exists, which suits a simulator of fictional patients, and no learning management system or student record system is connected either. An Anki integration for flashcards is listed as coming soon. Data moves only inside Oscer's own workflow for students and educators.
The platform runs in a browser, with individual student accounts and a separate institutional offer, so learners practice anywhere without a simulation lab. The privacy policy mentions outside service providers, account information and in app activity. Oscer's office is in Fitzroy, Melbourne, which says nothing about where the service runs.
No hosting provider, region or data location is named for chat logs, voice sessions or model requests, nothing describes whether one university's students are kept apart from another's, and no school can choose where its data rests.
Oscer Prime is sold directly to students and priced in USD. Promotional prices are US$6.99 for one month, US$34.99 for six months and US$59.99 for a year, shown against list prices of US$9.99, US$44.99 and US$69.99, and Prime unlocks every case, multiplayer practice, rubrics, the Condition Explorer and progress tracking. Students can sign up free with no card.
The terms separate prepaid periods from renewing subscriptions and explain cancellation through the account or the store that sold it. Universities speak with the team about curriculum matched cases, custom patients and tutor reporting, with institutional prices quoted separately.
Learners practicing history taking and diagnostic reasoning for OSCEs and US clinical skills exams are the primary users, across adult medicine cases, with nursing and paramedicine students among the testimonials. Tutors run curriculum matched cases and school wide OSCE style assessment and follow students through reporting. The University of Melbourne and the University of Sydney are named partners, and Oscer counts students from more than 150 universities in more than 35 countries. That reach supports the training setting. Whether practice on Oscer improves examination results or clinical performance, by discipline or learner level, has no measured outcome behind it.
Alternatives to Oscer
Other products that do the same kind of work as Oscer, and what each one does.
Xuron provides AI virtual humans for practicing healthcare conversations. Learners speak with simulated patients, caregivers or clinicians, receive feedback and repeat an activity.
PCS develops virtual patients, conversational simulation hardware and assessment tools for healthcare education. Spark combines spoken patient interviews with physical examination activities, simulated diagnostic and treatment decisions, and automated feedback.
MedVR Education provides virtual patient encounters and XR clinical skills training for universities, nursing programs and hospital educators.
Perceptors is an AI native learning platform for medical societies, universities, health systems and healthcare employers that want to turn their own approved clinical expertise into certified courses.
AI clinical simulation for nursing education and health system workforce training, delivered in immersive VR and screen-based formats.
SimX provides multiplayer VR simulation for healthcare education, combining virtual patients, clinical equipment and shared team encounters.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | BAA Tier | Implementation | Source |
|---|---|---|---|---|
|
US$6.99 for one month
$6.99 baseline
|
Oscer Prime costs US$6.99 for one month, US$34.99 for six months or US$59.99 for a year at promotional rates, against list prices of US$9.99, US$44.99 and US$69.99. Prime includes every case, solo and multiplayer practice, rubrics, the Condition Explorer and progress tools. | Student accounts, study activity and simulated sessions fall under the education privacy policy. No business associate agreement is offered, since the cases are simulated. | Students sign up directly. Universities arrange curriculum matched cases, custom patients and reporting with Oscer, with no institutional setup fee listed. | Vendor Published |
Free signup needs no card. The plans page shows promotional rates against higher list prices, and the terms separate renewing subscriptions from prepaid access.