FAQ
Questions, answered honestly.
Couldn't find what you need? Email us - a real human (one of us) replies within a day.
Getting startedCopied
Sessions are fully flexible - do a few quick stations in 15 minutes, or configure a full MMI circuit with longer read and answer times. It fits around your schedule, not the other way around.
A laptop or desktop with Chrome (recommended). A webcam is optional - only needed if you want facial analysis feedback. Mobile works too, though we recommend a laptop for the best experience. If mobile demand grows, a dedicated app may follow.
Yes - you'll need to sign up with Google or Discord, but it takes seconds and there's no credit card required. We link your analysis to your account so we can show your progress and trends over time.
About DeepmockCopied
Yes and no. At the base layer, we use frontier reasoning models, the same class of models everyone has access to. That part is not the product.
The product is the context we build around them. A generic model given a raw transcript has no idea what it is marking. Ours does, because every piece of practice is classified through a chain that narrows what "good" means at each step:
By the time your transcript reaches the model, it is not being asked "is this a good answer?" It is being asked whether this answer meets a defined standard, for this competency, in this speciality, against this rubric. That is the difference between generic feedback and feedback you can act on.
On top of that, we handle the parts a chat window cannot: timed station conditions, recording and transcription of your spoken answer, facial and delivery analysis, and tracking how you improve across sessions.
The product is the context we build around them. A generic model given a raw transcript has no idea what it is marking. Ours does, because every piece of practice is classified through a chain that narrows what "good" means at each step:
- Speciality: medicine, dentistry, nursing, specialty training. Each has its own expectations, regulator guidance, and language.
- Category: ethics, teamwork, motivation, communication, role-play. The competency domain the station is actually testing.
- Question: the specific prompt, its known traps, and what assessors are looking for in it.
- Rubric: the criteria your answer is scored against, with the weighting and descriptors for that question type.
By the time your transcript reaches the model, it is not being asked "is this a good answer?" It is being asked whether this answer meets a defined standard, for this competency, in this speciality, against this rubric. That is the difference between generic feedback and feedback you can act on.
On top of that, we handle the parts a chat window cannot: timed station conditions, recording and transcription of your spoken answer, facial and delivery analysis, and tracking how you improve across sessions.
Feedback is generated against the same kind of criteria a real assessor works from: a rubric for that specific question, with weighted criteria and descriptors, plus a model answer showing what full marks looks like. Your score is also placed against the distribution of every other attempt at that question, so a band means something relative to real candidates rather than an arbitrary number out of ten.
What it does reliably well:
What it cannot fully judge:
Use Deepmock for volume and iteration: the reps, the timing, and tightening your structure. Use real people for pressure and rapport. The candidates who improve fastest do both.
What it does reliably well:
- Whether you covered the points the question was actually testing, and what you missed
- Structure, signposting, and whether your reasoning is followable
- Timing, pacing, filler words, and delivery habits you cannot hear yourself doing
- Consistency: it marks your tenth attempt to the same standard as your first
What it cannot fully judge:
- Genuine rapport and the feel of the room
- How a specific panel on a specific day reads your presence
- Anything outside the transcript, your recording, and the rubric
Use Deepmock for volume and iteration: the reps, the timing, and tightening your structure. Use real people for pressure and rapport. The candidates who improve fastest do both.
Every question sits under a speciality and a competency category, and comes from one of two places:
Each question carries its own rubric and model answer, and both are regenerated automatically whenever the question itself changes, so the marking criteria never drift out of sync with the prompt you are answering. Your own questions get the same treatment as ours.
We do not publish leaked or confidential station material. Questions are written to the format and competencies of real interviews, not scraped from them.
- Written by the Deepmock team: built to match the formats and competency domains schools actually assess, drawing on published university guidance and assessor criteria. For the medicine speciality, questions are reviewed by a qualified doctor before they go live.
- Written by you: you can create your own questions at any time. Custom questions are private to your account, nobody else can see or practise them, and creating them is free.
Each question carries its own rubric and model answer, and both are regenerated automatically whenever the question itself changes, so the marking criteria never drift out of sync with the prompt you are answering. Your own questions get the same treatment as ours.
We do not publish leaked or confidential station material. Questions are written to the format and competencies of real interviews, not scraped from them.
Your recordings are private by default. Nobody else can see a practice session unless you deliberately share it.
See our Privacy Policy for the complete detail.
- Storage: recordings are held in private storage tied to your account and your region, and are never used to train models.
- Sharing is opt-in and granular: if you create a share link, you choose whether the viewer can see your score, leave comments, or watch the video at all. Links expire by default, and you can revoke one at any time.
- Facial analysis is opt-in: delivery and expression analysis only runs after you explicitly consent, and you can practise without it.
- Region isolation: your data lives in your region's database and never crosses into another.
- Deletion: you can delete an individual practice session, or request full account deletion, which removes your recordings along with your account.
See our Privacy Policy for the complete detail.
Deepmock supports Australia and the UK, each with its own content and question bank, and covers both major formats:
Many universities have a saved format preset, with station count and timings matching that school's published circuit, so you can practise under the exact conditions you will face. If your university is not listed, you can still build a series with custom timings and use the same question bank; the competencies assessed are broadly consistent across schools.
Coverage is not limited to medicine. Each speciality carries its own categories, questions, and marking context, so the standard you are assessed against matches the field you are applying to.
- MMI series: multi-station practice with timed preparation and answer periods, run back to back the way a real circuit runs.
- Solo practice: a single question at a time, for drilling one competency or one weak spot.
Many universities have a saved format preset, with station count and timings matching that school's published circuit, so you can practise under the exact conditions you will face. If your university is not listed, you can still build a series with custom timings and use the same question bank; the competencies assessed are broadly consistent across schools.
Coverage is not limited to medicine. Each speciality carries its own categories, questions, and marking context, so the standard you are assessed against matches the field you are applying to.
Privacy & recordingsCopied
Only you, by default. If you want to share for peer review, you can - it gets its own link that you can expire at any time. Nothing is ever shared publicly, and we don't train AI models on your videos.
Yes - you can delete your recordings any time from your account. Deleting a recording removes the video, but any linked analysis stays.
No, never have, never will. AI feedback runs at request time only - your recordings are never used as training data. Our models are pretrained and our prompts and rubrics are reviewed by clinicians. Your sessions are yours.
Pricing & plansCopied
No, and we won't. Free gives you 8 AI analyses a week, forever. Paid plans are one-time charges - they auto-convert back to Free when the window ends. No card stored, no forgetting to cancel a subscription. You pay for what you want, that's it.
- Cram - 7 days
- Focused - 30 days
- Season - 90 days
Yes. Unlimited practice and peer review, always free. We limit AI analyses to 8 per week on the free plan. We make money from our one-time plans - Cram, Focused, and Season - never recurring billing.
If price is a barrier, email support@deepmock.io and we'll sort something out if possible. Students in financial need matter to us - we never want cost to be the reason you can't do your best.
All three unlock every feature with unlimited usage. Just pick your window:
- Cram - 7 days. For those who found us late and want that final push before their interview.
- Focused - 30 days. Our most popular. Built for the 2-4 weeks of notice most students get.
- Season - 90 days. For applicants juggling several universities and rounds.
Never. Cram, Focused, and Season are one-time payments. When the window ends you drop back to Free with all your data intact. We don't store your card after the purchase, so we couldn't charge you again even if we wanted to.
Scenarios & feedbackCopied
We don't replicate exact MMI questions - that would breach university NDAs. Instead, we work with clinicians who have been part of the interview examination process to create scenarios of the highest quality, mirroring the real format, structure, and difficulty.
No, and we don't claim it. A good coach is irreplaceable. What Deepmock does is the volume - the reps that make your tutor session worth the money. Many students arrive at their tutor better prepared because of it. We're also working on tutor integrations so coaches can assign and review sessions directly in the product. If you're a tutor interested in partnering, reach out at business@deepmock.io.