Diagnosica
Comparisons

Sketchy Cases review, hands-on from a one-case trial

Mostafa Ibrahim9 min read
Sketchy Cases review, hands-on from a one-case trial
We build Diagnosica, a competing AI clinical simulator. Nobody paid us or gave us anything for this review. We paid for our own Sketchy trial. Sketchy facts were checked on its live pages in October 2026, and we link them later in the post.

If you search for a Sketchy Cases review, you mostly land on Sketchy’s own pages or partner content. We wanted a plain, hands-on writeup from someone who builds a simulator too, with scope and limits stated up front.

Our session was one run on August 30, 2026 in desktop Chrome, signed in on a paid trial. We opened a single case, a 2-year-old with vomiting and diarrhea, and played the first four of ten steps: History, Physical Exam, Presentation, Differential Diagnosis. We captured 54 screenshots.

We did not reach steps five through ten, so we won’t judge scoring, feedback, debrief, or teaching points. Where we reference those areas, we’ll attribute Sketchy’s claims to its live pages and keep them separate from what we actually saw.

Sketchy Cases review, hands-on from a one-case trial

Is Sketchy Cases worth it?

For a US student who wants long, structured cases with an attending guiding you step by step after the history, yes, it is worth a look. The history-taking is genuinely fun. Downsides are no visible free tier, a trial that needs a saved payment method, iOS-only in the US, and no habit features in our August 30, 2026 session.

Scope first: we ran one case on August 30, 2026 in desktop Chrome, signed in on a paid trial. We played 4 of 10 steps in a 2-year-old with vomiting and diarrhea, so we cannot speak to anything beyond Differential Diagnosis in this review.

What Sketchy Cases is, and who it's for

Sketchy Cases is a virtual patient platform where you type or talk through a five-step workflow, History & Physical, Differential, Orders & Results, Management, Feedback. All of that is from Sketchy’s materials on Sketchy's own Cases page.

They market an in-case supervisor, the attending, who “corrects your reasoning in real time, with high-yield teaching points and MCQs built in.” That is their claim, and we did not reach those steps in our session.

They list “405+ cases across every core rotation and setting” and show program tracks for Medical, PA, NP, and MCAT on the same page. The pitch sits squarely at students heading into or on rotations, and those gearing up for exams.

With speech-to-text, you can talk to the virtual patient and “present your SOAP out loud,” which fits learners who prefer to rehearse oral presentations.

There is a games shelf framed as “The study break that still counts,” with Mini Cases, Board blitz, Bugs & Drugs, Stat Read, DxDash, GuesSTI, Path-o-Finder, and Bootcamp as samplers.

Sketchy also cites outcomes from its own surveys, 92% of M3s said Cases helped them feel more prepared for rotations, and 81% recommend. Those are Sketchy’s marketing figures with no sample size stated.

Diagram of Sketchy Cases' ten case steps, with the four we played marked and the six we did not

Inside a case: ten steps and three panes

In this Sketchy Cases review, our session opened to a full case view that felt structured and busy in a good way. The patient is an illustrated card with a mic for speech-to-text, and the library badge said 25 to 35 minutes for a full run.

Across the top sit ten tabs, each a step: History, Physical Exam, Presentation, Differential Diagnosis, Orders, Results, Management, Course of Care, Case Summary, Teaching Points. As you progress, completed tabs turn green. Sketchy’s marketing groups the ten into five headers, “History & Physical”, “Differential”, “Orders & Results”, “Management”, “Feedback.”

The layout is three panes. Left is the Chart with a patient thumbnail, vitals, an autosaving Notes pad, and your earned key findings grouped by step. Center is the conversation, with counters for findings and messages. Right is a persistent DDx Builder that says, “Drag to rank by likelihood.”

There are two AI characters. You take the history from the patient’s mother, then the attending, Elizabeth Netsu, MD, Pediatrics, runs the physical exam and every step after. An age-appropriate historian and a named attending is more than most simulators offer.

Our case was “Heidi Myers, a 2-year-old girl with vomiting and diarrhea.” Vitals on load were BP 88/54, HR 128, RR 26, SpO2 99 percent, T 100.8 F.

Voice output is built with ElevenLabs voices, designed for characters like hesitant, warm, assertive, and confused, per the ElevenLabs case study. Nothing auto-played for us, so we stayed text-in and silent on output.

The best part: key findings you earn as you go

This is Sketchy Cases at its best. The history opens at 0 of 7 key findings, you ask one well-aimed question, and three gold chips pop at once. The counter jumps, the chart fills, and you feel the case move under your hands.

Our session was the 2-year-old with vomiting and diarrhea. We asked about frequency and blood, and three findings unlocked at once as gold Key Finding chips. The counter went to 3 of 7, and those chips landed in the Chart on the left, grouped by step, next to vitals, a patient thumbnail, and an autosaving Notes pad.

The DDx Builder is persistent, and you can drag to rank by likelihood. When the attending asks for your top three, that ranked list is already there. If you want a refresher, here's our guide to ranking a differential properly.

Worked breakdown of the history step in our session, showing key findings counter, message budget, free hint and a guard-rail

The rules: message budgets, free hints, a graded presentation

Sketchy shows message budgets as a rule of the game. History gives you 40 messages, then physical exam and later steps give you 20, with a live counter on screen.

We clicked the free hint once. It nudged a direction, “ask about recent travel,” not the answer, did not spend a message, and then asked, “Was this hint helpful?” Following it unlocked two more key findings.

When we tried to press Continue at 5 of 7 findings, we got a guard-rail warning that more key findings remain and the conversation would end if we moved on. Sensible friction before you miss something important.

Physical exam in our session was free-text maneuver requests against 0 of 6 findings. Asking to assess hydration returned capillary refill about 3 seconds, dry mucous membranes, and decreased tears, with an incidental eczema finding. The eczema distractor makes sense for pediatrics, and it rewards a clean problem representation.

The presentation step felt true to rounds. The attending asks you to summarize the patient, you can type or speak, and it is graded live, with the on-screen line, “You hit everything I wanted to hear.” A “Problem Representation” chip highlights the core. This rehearses what attendings expect on rounds, and links naturally to doing a tight patient presentation. As a second genuine strength, that step stands out.

What we didn't test, and what that means

We ran one session on August 30, 2026 in desktop Chrome on a paid trial. We played a single case through four of its ten steps, stopping after History, Physical Exam, Presentation, and Differential Diagnosis.

We didn’t reach Orders, Results, Management, Course of Care, Case Summary, or Teaching Points.

Because we stopped early, scoring, feedback, and any peer comparison stay unverified. Same for how labs and imaging are shown, the quality of teaching points, the mini games, and voice output. We never triggered voice output, so its usefulness is unknown to us.

Sketchy says, “Your virtual attending corrects your reasoning in real time, with high-yield teaching points and MCQs built in.” That is their claim, and we couldn’t check it within the steps we played.

If you’re deciding, use the trial to test exactly those later pieces. We could be wrong about this, but your call probably turns on them.

What's missing: no streak, no daily habit, no leaderboard

We were signed in on August 30, 2026, and saw no streak, no daily goal, no points, and no leaderboard anywhere in navigation, home, or the case library. The games shelf was there, but in our session it seemed to count toward nothing we could see.

Full cases in the library are badged 25 to 35 minutes, which is honest about the lift but leaves no obvious small daily unit. We did not see anything like a one-minute check-in or two-minute case.

Sketchy’s public Cases page lists Mini Cases, Board blitz, Bugs & Drugs, Stat Read, DxDash, GuesSTI, Path-o-Finder, and Bootcamp, framed as “The study break that still counts.” We did not see those gamified bits tie into a habit loop in-app during our session.

This matters if you study in short bursts or rely on streak psychology. If you want that Duolingo feel, see a Duolingo for med students. It may change, but this is what we saw on August 30, 2026.

Price, and who should choose Sketchy Cases

Cases-only is $19.99 a month or $119.99 a year as of October 2026. All Access runs $349.99 for 6 months, $449.99 for 12, or $649.99 for 24. You get a 7-day trial with a saved payment method, subscriptions auto-renew, and sales are final.

You can confirm the Cases-only and All Access tiers on the Cases pricing page, which also shows these rates as of October 2026. We didn’t see a free tier described there, only free-trial language.

The trial needs a payment method, and Cases plans auto-renew unless you cancel. Their refund page states, in full, "All subscription purchases are final", so set a reminder if you’re testing it near exams.

If you’re scanning the field before committing, here are clinical reasoning apps ranked.

Choose Sketchy Cases if:

  • You want long, structured cases where the attending steps in at each stage of reasoning.
  • You already subscribe to Sketchy’s Medical, MCAT, PA, or NP programs and want cases in the same place.
  • You want presentation practice inside a single case rather than quick one-liners.
  • You mostly study on desktop or iPhone in the US.

If it's not the fit: where Diagnosica differs

Diagnosica has no attending and no debrief. Sketchy’s attending-led steps feel deeper than our encounter, and we don’t have a key-findings counter or a presentation step. You do get a scorecard after every case here, but that’s not a debrief.

If you’re also sizing up chat-first tools, here’s how Neural Consult compares. Broader view of study options, we keep a short list of AI tools for med students.

  • Patient. Sketchy Cases: static illustration. Diagnosica: a video avatar you speak to or type to.
  • Trying it. Sketchy Cases: free trial, no free tier described. Diagnosica: free tier, no card needed.
  • Phones. Sketchy Cases: iOS only, US region, no Android. Diagnosica: iPhone, Android, and web.
  • Short daily play. Sketchy Cases: none seen in our session. Diagnosica: a two-minute quick case with daily streaks and weekly leagues.

Diagnosica doesn't give advice about real patients, and it's not a diagnostic system. Treat every output as educational. Try a case free

Run your first case tonight. No demo call, no card. Sign up, pick a case, and you are talking to a patient within a couple of minutes. Start free on Diagnosica

Sketchy Cases facts were checked on Sketchy's live pages on October 9, 2026; hands-on observations are from one session on August 30, 2026. Educational use only, not medical advice.