Academic judgment, made visible.
Rocketproof is an AI-powered assessment and simulation platform for higher education. Students defend their work in oral exams and practice inside realistic AI role-play. Instructors build both with AI built into the platform, then review the transcript, score the rubric, and release the grade themselves.
Instructor? Start with Faculty Pro.
Rocketproof is an assessment and simulation platform for higher education. Students defend what they know and practice inside realistic scenarios. Instructors create both with AI built into the platform, then review the same kind of evidence for each.
Assess
Oral exams, defenses of papers and presentations, reading checks, and cumulative assessments. Questions come from the student’s own work and every session leaves a transcript.
Simulate
The AI plays a patient, client, interviewer or stakeholder that you define. Students talk it through, make decisions, and are assessed on how they handled the situation.
Build with AI
AI assistants built into Rocketproof draft assignments, simulations, rubrics and question sets, and analyze course results. They work inside your role and your courses.
The shared review layer
A role-play counts in a course for the same reason an oral defense does: there is a record, a rubric, and a person who decides.
Transcript
Recording
Rubric scores
AI analysis
Faculty judgment
You build the experience
Describe it to the built-in AI or set it up yourself: an oral defense, a role-play, a scenario.
Students take part
Defend a paper, a deck or a reading, or talk with an AI persona you defined, by voice or text.
You review the evidence
Transcript, recording, and rubric score in one place. Release when ready.
Runs inside your LMS
Rocketproof is an LTI 1.3 tool with Assignment and Grade Services and Names and Role Provisioning Services. Canvas is the platform we have verified end to end, from registration through a graded launch. If your institution runs a different LTI 1.3 platform, talk to us before you plan a term around it. LMS integration and platform status.
No video to press play on, no imagination required. This is the shape of an oral defense: a question generated from the student's own work, a spoken answer, and a rubric score instructors can stand behind.
Question 1 · generated from the student's own submission
Your methods section says you used a paired t-test. Why was that the right choice for this design?
Spoken response
Because every participant was measured twice, before and after the intervention, so the samples are dependent, not independent…
You define the situation, the character and the goal. The AI plays that character and responds to what the student says. Use it for practice that stays out of the gradebook, or as a graded assessment scored against your rubric.
AI patient · chest discomfort
Student goal: Take a focused history and recommend a next step.
It's probably nothing. I just get this tightness when I climb the stairs at work.
Thanks for telling me. How long has the tightness lasted each time, and does it ease when you stop?
Students should interview a patient with chest discomfort who downplays it. Twelve minutes, graded.
Draft ready: persona and history, student goal, twelve-minute limit, four rubric criteria. Nothing is published.
Rocketproof includes AI agents that work on the same courses, assignments, rubrics and student evidence the platform already holds. Describe an assessment or a simulation and one drafts it. Ask how a course is going and another answers from your own results. They draft and advise. You save, publish and grade.
A walk-through of the platform end to end, about twenty-six minutes long: how oral defenses work, from the question plan Rocketproof builds, through a recorded session, to the instructor review surface where the rubric gets scored.
Doctoral programs have used the oral defense for a century because a written artifact can be produced without being understood. Rocketproof brings that method to an undergraduate section, and gives the instructor one review surface instead of a term of office hours.
Every paper upload, presentation delivery, reading response and simulated conversation becomes a defensible evidence trail. Instructors review the record. Students know the standard. Institutions keep the rigor.
Submission-only grading and honor pledges can't show you what a student understands. A recorded, source-aware oral defense can.
| Capability | Rocketproof | Paper only | Honor pledge |
|---|---|---|---|
Practice in realistic situations | |||
Evidence of actual understanding | |||
Structured instructor review surface | partial | ||
Recorded transcript + audio/video | |||
Source-aware questioning | |||
Defensible audit trail | |||
Integrity monitoring controls | |||
Rubric-aligned scoring | manual | ||
Runs at full section size | |||
Nothing to schedule |
Want a direct evaluation of platform alternatives? Compare AI oral exam platform architectures and DIY prompting.
Scalable oral exams implementation guidePapers, decks, reading passages, and scenario prompts all shape the question path before the session starts.
A 40-page thesis generates 12 questions tied to its argument structure and cited sources.
Transcripts, delivery clips, identity checks, and environment scan artifacts stay organized inside one review surface.
Instructors open one panel and see a 14-minute recording, full transcript, identity check result, and rubric scores side by side.
Enable only what matters: ID checks, environment scans, prompt leakage review, and anomaly alerts.
Enable ID check and environment scan for high-stakes finals; skip them entirely for low-stakes reading checks.
Source summaries, strengths, weak spots, hesitation patterns, and rubric scoring all in one review panel.
The AI flags "student hesitated on methodology questions" with a follow-up prompt the instructor can use immediately.
Write the situation, the character, the goal and the limits. The AI holds the persona and reacts to what the student says.
A nursing instructor writes an anxious patient who downplays chest discomfort, and students take the history by voice.
Draft assignments, simulations, rubrics and question sets in conversation, and ask questions about your results.
Describe a twelve-minute patient interview and get a draft persona, goal and rubric to edit before anything is published.
They are grouped by what you want a student to do, not by what they hand in. Every one of them produces the same structured evidence trail, so review stays consistent whichever you choose.
The student hands in work, then answers for it.
Students upload a paper or document, then answer oral-defense questions about it.
Students upload a slide deck, and the AI asks questions about the presentation itself.
The student shows what they know, without submitting an artifact first.
You provide the reading, and students defend their comprehension of the assigned material.
The AI examines the student, asking follow-up questions to probe understanding.
A live, open conversation with the AI about the student’s work or thinking.
A replacement attempt for a student who missed or must retake an assessment.
The student acts inside a situation, role, or conversation.
A conversation the instructor steers with a greeting, a scenario, a source and a goal, so the student is working toward something specific rather than talking freely.
Students work inside a situation you describe, responding as it develops.
The student does the questioning. Rocketproof plays a patient, client, witness or source, and the student has to get what they need out of the conversation.
A spoken conversation in the target language, assessed for fluency and accuracy.
The student takes a position, or evaluates another student’s work.
Students evaluate each other’s work and defend their assessment.
The AI argues a position and the student must respond and hold their own.
The student draws across several assignments, sources, or a whole term.
A final assessment built for each student from their own prior work in the course, rather than one paper every student answers.
The fast version of the overview, focused on the core product flow: watch a paper turn into a question plan, a student answer it, and the instructor review surface when the session lands.
Runs about seventeen minutes, no signup required.
Students upload the paper. Rocketproof builds a targeted defense plan. Instructors review the source summary and Q&A record together.
Deck upload, delivery recording, transcript capture, and deck-aware Q&A all in a single workflow.
Reading Assessment checks that the assigned text was read and understood. Set the discipline to Humanities and Literature and the same assignment presses on interpretation, textual evidence, and competing readings instead.
Write the situation and the character. The AI plays it, by voice or text, and the student works toward the goal you set. Practice stays out of the gradebook. Graded runs are scored against your rubric.
The assistants inside Rocketproof draft assignments, simulations, rubrics, question sets and course structures from a conversation. They can also answer questions about your results. They draft and advise, and you decide.
Choose the assignment type or describe it to the built-in AI. Attach the source, or write the persona and scenario.
Students upload, complete checks, then defend their work or talk with the AI persona.
Questions, transcripts, recordings, and signals captured automatically.
Read the summary, inspect the transcript, confirm the rubric, release what you approve.
See how it would work in your courses.
Start with a focused group of courses or faculty, and decide together whether to expand.
From the moment an instructor creates an assignment to the moment they release feedback, Rocketproof carries the full evidence trail. Three real product views, in order, with no audio, a silent demo reel.
Instructors create an assignment, choose settings, select an assignment type, and assign it to students.
Students receive the assignment, respond through the guided workflow, and submit their work for review.
Instructors review the student’s submission, transcript, AI-assisted analysis, and scoring support.
The questions your IT and compliance offices will ask, answered in plain language, not buried in a security PDF.
Student records are processed under your institution’s direction, consistent with FERPA. The full policy is public. No NDA required to read it.
Read the privacy policyStudents launch from the course they are already in, with no second password to remember. Grades pass back to the gradebook, and the roster syncs. Verified end to end with Canvas.
LMS integration and platform statusRecordings, transcripts, and integrity artifacts are stored behind signed access links, scoped by your retention settings.
Extended time, audio-only mode, typed responses, and question pacing, set per student or for the whole course. WCAG 2.1 AA is the standard the interface is built and audited against.
Rocketproof is an AI-powered assessment and simulation platform for higher education. Instructors use it to run oral assessments of student work, to create realistic AI role-play and scenario simulations, and to design and manage both with AI assistants built into the platform. Every session leaves a transcript, rubric-aligned evidence and an instructor review step before a grade is released.
Read why Rocketproof existsRocketproof does three things in one platform. It assesses understanding through oral exams and oral defenses. It simulates professional situations through AI role-play, so students can practice or be assessed in realistic scenarios. It includes AI agents that help instructors build assignments, rubrics and courses and analyze results. Instructors keep the final say on every grade.
See the three capabilitiesYes. Rocketproof supports AI role-play in which the AI holds a persona you define, such as a patient, client, interviewer or stakeholder, and responds to what the student actually says. Students can take part by voice or by text. Role-play can be used for practice that does not count toward a grade, or as a graded assessment with a rubric.
AI simulations and role-playYes. Rocketproof runs two kinds of simulation. A conversational simulation puts the student in a dialogue with an AI persona inside a situation you describe. A branching scenario presents a situation that changes with each decision the student makes and asks them to justify those decisions. Both produce a transcript for instructor review.
How Rocketproof simulations workYes. Rocketproof conducts oral exams and oral defenses. The AI examines the student, asks follow-up questions, and produces a transcript and draft rubric scores for the instructor. Students can answer aloud, in a live voice conversation, or by typing, and approved accommodations such as extended time and audio-only are supported.
AI oral exams and oral defenseYes. For a role-play or guided conversation, the instructor gives the AI a name, a role, instructions for how it behaves and what it may reveal, a greeting, the student’s goal, and optionally a voice and an avatar image. Instructors can write these directly or ask the built-in assistant to draft them.
Build a persona and scenarioYes. Rocketproof includes AI assistants for instructors that draft assignments, simulations, rubrics, question sets and course structures, and that answer questions about course and submission results. Students have a per-course Course Guide and a Study Coach. The assistants draft and advise. An instructor saves, publishes and grades. Rocketproof does not currently offer an MCP server or a public API for connecting external assistants.
Built-in AI agents in RocketproofYes. An assignment can be marked as practice. Practice attempts are kept out of the gradebook and out of grade passback, so a student can rehearse an oral defense, a negotiation or a patient interview as many times as the instructor allows.
Practice with AI simulationsYou decide, and the assignment builder shows you the estimate before you publish. It is calculated from the question count, the think time, and the answer window you set. A five-question defense with a two-minute answer window estimates at about fourteen minutes.
The recording, the transcript, the rubric scores with their rationales, the source summary, and any integrity artifacts the assignment enabled. Everything sits in private storage reached through signed links, and your institution sets the retention window.
Security and data practicesYes. Instructors can upload or recreate their rubric criteria inside Rocketproof, or ask the built-in assistant to draft one. The platform maps rubric dimensions to session questions and scoring.
Yes. Text Conversation is available across supported assessment workflows and requires no camera or microphone unless an instructor independently enables an integrity check. Audio-only and camera-off options remain available through approved accommodations for spoken assessments.
Yes. Extended time multipliers, audio-only mode, and question pacing settings are all configurable per-student or course-wide through accommodations settings.
Accessibility statementYes. Rocketproof is an LTI 1.3 tool with deep linking, grade passback through Assignment and Grade Services, and roster sync through Names and Role Provisioning Services. Canvas is the platform we have verified end to end, from registration through a graded launch.
LMS integration and setupYes. Use the Reading Assessment type and set the discipline to Humanities and Literature. The questioning then presses on close reading, textual evidence, and competing interpretations, and the instructor sets how deep it goes.
AI oral exams and oral defenseSessions are recoverable. Students can resume from the last completed question within the session window. All captured data before disconnection is preserved.
Pricing scales with the course. Student seats are priced per semester pass, and department or institution licenses add volume discounts and centralized billing. Talk with us for a quote that fits your sections.
Pricing and plansNo. Rocketproof conducts the conversation and prepares draft analysis. The instructor reviews the transcript, adjusts the scores and decides what is released. The built-in assistants draft and advise, and they do not grade.
Request a pilot to try it with a focused group of courses, ask for a demonstration, or explore the interactive demo first.
