Practice the interview before it counts.
ClinSim trains clinicians and students with voice-native simulated patients. Configurable personas present realistic symptom interviews, and every session ends with an automatic debrief scored against standard clinical rubrics.
In development. A training simulator for professionals, never a source of care.
Automatic debrief
The problem
Standardized patients are excellent teachers. They are also scarce.
Clinical interviewing and symptom-rating skills are still taught almost entirely through live sessions with trained actors. The format works, but it cannot scale to the number of reps a trainee actually needs.
Expensive by design
Every session needs a trained actor and a supervising clinician. Budgets cap most trainees at a handful of practice interviews per year.
Bound to the calendar
Rooms, actors, and faculty schedules all have to line up. Practice happens when the schedule allows, rarely when the trainee is ready to learn.
Inconsistent feedback
Two observers can score the same interview differently. Trainees get feedback that varies with who happened to be in the room.
The same bottleneck reaches into clinical research: rater training for trials depends on scarce calibration sessions, and scoring drift between raters adds noise that studies pay for in statistical power.
How it works
Choose a case, hold the interview, read the debrief.
A full practice loop in one sitting, as many times as it takes.
Pick an instrument and difficulty
Select a standard clinical instrument and a difficulty level. ClinSim loads a synthetic persona configured to a target severity profile, with ground-truth item scores set before you say a word.
Conduct a spoken interview
Talk with the persona in natural voice, the way you would sit with a patient. It answers in character, holds its story consistent, and presents symptoms at the configured severity.
Get a rubric-based debrief
The session is scored automatically against the instrument's rubric: which items you covered, how your technique held up, and how close your ratings came to the persona's ground truth.
What's inside
Built around the practice loop
Persona library
Synthetic patients spanning presentations and severity levels, each with a consistent history and ground-truth scores. New personas are authored and checked for fidelity, never drawn from patient data.
Voice-native sessions
Interviews are spoken, in real time. Pacing, silences, and follow-up questions all matter, so the practice transfers to the room where it counts.
Rubric scoring
Debriefs follow the instrument's own rubric: item coverage, interview technique, and your scoring accuracy measured against the persona's configured ground truth.
Cohort dashboards Planned
For training programs and rater-training teams: track coverage, technique, and calibration across a cohort, and spot drift before it reaches a study.
Where we are
In development, built in the open
ClinSim is an early-stage project. The persona and scoring loop is being built for a first instrument now, with voice sessions and pilots with training cohorts to follow. If you teach clinical interviewing, run rater training, or are preparing for OSCEs yourself, we would like to hear from you.
Join the waitlistOne email when early access opens. Nothing else.
FAQ
Common questions
Is this therapy?
No. ClinSim is a training simulator for clinicians and students. The "patients" are synthetic personas, and the only person being evaluated is the trainee. It does not provide therapy, diagnosis, or any form of care, and it is never used with real patients.
Who is ClinSim for?
Medical and psychology training programs, rater-training departments at research organizations, and individual trainees preparing for clinical interviews or OSCEs. Access starts with individual trainees; program and site licenses come later.
Where do the personas come from?
Every persona is authored synthetically, specified by a severity profile with ground-truth item scores, and checked for consistency before release. No patient data is used or collected, at any stage.
How is my interview scored?
An automated pipeline evaluates the session transcript against the instrument's published rubric: which items you covered, how your questioning technique held up, and how closely your ratings matched the persona's configured ground truth. Scoring quality is checked in recurring reliability runs.
Is ClinSim a medical device?
No. ClinSim is educational software for professional training. It makes no claims about diagnosing, treating, or managing any condition in real patients.
When can I use it?
ClinSim is in active development. Join the waitlist and you will get one email when early access opens.