VivaCoach Prep — User Guide

Your systematic warm-up for the defense. This guide walks you through how to use it; it doubles as
the getting-started guide for beta testers.

You're in the beta. New here? Start with the one-page Beta Welcome,
then come back for the details. The beta-specific parts — your data, usage limits, giving
feedback, what's still rough, and your tester mission — are near the bottom.

What this is (in one breath)

VivaCoach Prep turns "I hope they don't ask something I can't answer" into a checklist you can
actually work through. It lays your thesis out as a grid: your research questions and claims
down the side, your committee members across the top. Each cell is "what would this examiner
probe about this part of my thesis?" You rehearse cell by cell, mark what's solid and what's
shaky, and walk in with systematic coverage instead of open-ended worry.

It's a rehearsal room, not a fortune-teller — the questions are well-grounded anticipations, not
predictions of the exact wording you'll get.

The mental model: Know it → Say it → Defend it

Everything is organized around three rungs you climb per cell, in parallel (you don't have to
finish one before starting the next):

1. Know it — Can you write a solid answer? (typed practice)
2. Say it — Can you deliver it aloud, no notes? (spoken practice)
3. Defend it — Can you hold the whole thesis together under pressure? (the full mock viva)

The top of the screen always shows where you are: Setup → Prepare → Defend.

1. Setup (once)

1. Add your thesis. Upload your PDF (or explore a demo). The app extracts your research questions and claims.
2. Confirm. Check the extracted research questions/claims and fix anything that's off.
3. Set your committee. Add your real examiners (it builds a short profile from public info), and optionally add a Synthetic critic — an AI archetype (e.g. the Blind Spot Critic) that probes the angles a tough examiner might.

Your work saves automatically in your browser. You can leave any time and pick up where you left
off. (Demo theses are for exploring — changes there aren't kept; upload your own to save work.)

To edit setup later, click Setup in the top bar; when you're done, hit ✓ Done to drop
right back where you were.

2. Prepare — work the grid

Home is your hub — where you stand and the single best thing to do next. For the full picture, open the Coverage Map: the grid of every research question × examiner cell, colored by how ready you are.

• The Know it / Say it lens (toggle at the top) recolors the grid to show your written prep or your spoken prep.
• Cell colors / bands: Solid (you can defend it) · Wobbly (getting there) · Weak (needs work) · Not yet (untouched).
• Click any cell to read what that examiner is likely to probe, then practice it.

Practicing a cell: you get the cell's questions one at a time, each headed with the examiner asking it and why — drawn from their work — so you're always rehearsing a specific person's line of attack, not a generic prompt. Type your answer (Know it) or
tap 🎙 Speak to answer aloud (Say it). Submit, and the coach scores it on coverage, depth,
framing, tells you what was missing, and offers a stronger opening. A low score isn't a
verdict — it points at one specific, fixable thing. Tighten that and take another pass.

Every answer saves the moment it's scored, and the cell remembers which questions you've
answered — leave whenever you like, and opening the cell again picks up at the first question you
haven't done. Finish & review closes the run with a recap; Discard this run throws the
run's answers away if you'd rather it not count.

While answering you can use:

• Hint — what this examiner is really after + territory to cover.
• Suggest — a model answer (use it as a starting point, or flag it if it misrepresents your work).
• Challenge — push back on the question itself when it rests on a faulty premise. The coach responds with a verdict: it either upholds the question (and tells you why it stands) or concedes and rewrites the question — so challenging is a real exchange, not a dead end.

Whole-thesis band: at the top of the grid is a set of 13 canonical defense questions every
viva asks — "Summarize your thesis," "What's your original contribution," "What are the limitations,"
"How did your research questions evolve," "What recent work bears on your thesis," "What would you do
differently / what's next?" These cut across the whole thesis (not any single claim) — real vivas are
full of them — so practice them too.

Next up: Home surfaces a ranked queue — the single most useful thing to practice next
(your weakest, your flagged, the ones you haven't touched). One tap drops you straight into it.

Mark as you go: flag cells that need work; mark cells Ready when you can defend them. The
coverage readout up top tells you, in plain language, how much of your thesis surface you've
covered.

What the labels mean (we keep this honest): a cell's confidence chip (High / Medium / Low) is
match strength — how strongly that examiner's specialty overlaps the topic — not a promise the
question will be asked. Open a cell and you'll also see an evidence marker (◌ Exploratory): every
anticipation today is AI-generated and not yet validated against real defenses. Treat the grid as a
well-aimed study map, not an answer key.

3. Say it — rehearse aloud

Saying an answer is what turns a draft you know into an answer you can deliver. Switch the lens
to Say it, and Home shows "best to say next." In any cell, choose 🎙 Say it (aloud),
tap Speak, and talk — your words fill the answer box (edit before submitting if you like).
It's scored on delivery, lenient on exact phrasing.

First time speaking? Your browser asks for microphone permission — allow it. Works best
in Chrome; if Speak doesn't respond, check the mic isn't blocked in the address-bar site
settings. Your words are transcribed to text and scored on that — in this beta we don't keep
the audio. Worth a 10-second test the first time — tap Speak and say a word — so a blocked
mic surfaces before a timed mock viva, not in the middle of one.

4. Defend it — the mock viva

Click Defend in the top bar to open your Defend status page: your past mock-viva runs,
your last score and trend, and the buttons to start a new one.

Start a mock viva — your whole committee takes turns, end to end, unaided. Like a real defense it
opens with "Summarize your thesis" and closes with the big whole-thesis questions (limitations,
what you'd do differently / what's next), with your committee's claim-specific probes in between. By
default it runs realistic: no scores until the finish, closest to the real room. (You can switch to
coached — scored after each answer — when you set it up.) Answer aloud or typed — either way
it's scored as spoken delivery: a viva is an oral exam, so there's no written-mode toggle
mid-run (typing is always fine; it's judged as what you'd say).

• Pause — resume later: leave any time; your progress is saved and a Resume banner waits for you on the Defend page.
• Finish & review: ends the run and shows your recap — per-question scores, your strongest and weakest moments, and the single thing to work on next.

Build your own viva: once you've saved some questions (below), assemble a custom run from
exactly the ones you want, pick coached or realistic, and run them in order.

5. Saving questions (★)

The questions that stump you are the most valuable to revisit. Tap ★ Save on any question —
during a run, in the score panel, or on the recap/review afterward. Saving stores the question
and your answer, so later you can see both. Saved questions:

• become practiceable again from their cell (they join that cell's question pool),
• collect on the Home ★ Saved questions shelf (with your answer), and
• can be assembled into a build-your-own viva.

The ★ is a toggle — tap again to unsave. (You can save a hard question before you have a good
answer, to come back to it.)

6. Reviewing & tracking progress

• After any run, "Full review →" shows the complete transcript with scores, feedback, and copy/save buttons per question.
• Past sessions keeps every run — reopen any to review.
• Progress & trends shows how your coverage and readiness are moving over time.

Beta: your data & privacy

Your thesis is confidential research, and it's treated that way:

• Encrypted at rest, stored with per-user database access controls — other testers can't see your work, and neither can your advisor.
• To generate questions and score answers, your thesis text is processed by our AI provider under terms that prohibit training use — your thesis is never used to train AI models.
• You're in control: delete your account and all your data at any time, in the app or by email. We send login and service notices only — no marketing.
• We'll only ever email you about the beta itself.

At signup you'll confirm you're 18+ and agree to the Terms and Privacy Policy. There's also an
optional per-session checkbox to let an anonymized copy of a session help us improve scoring —
entirely your choice, and you can withdraw it for unused sessions.

Beta: usage limits

The AI calls behind questions and scoring cost real money, so each tester has a weekly usage
allowance. It's set generously for normal prep — most people won't notice it. You'll see a
friendly banner as you approach it, and AI features pause if you hit the ceiling (your saved
work is untouched; the limit resets weekly). If you bump into it during honest prep, tell us —
that's useful signal, and we can raise your cap.

Beta: giving feedback (this is the whole point)

You're here to find the rough edges — please be blunt. Three ways, fastest first:

1. 👍 / 👎 on a score — was that score fair? One tap, with an optional one-line comment. This is the single most valuable signal you can give.
2. The Feedback button (bottom-left, in the app) — leave a note or report a bug anytime. Bug reports attach the basics (what screen, app version) automatically, so you don't have to reconstruct what happened.
3. Email me directly — mbidun@gmail.com — for anything bigger, or if you're stuck. You're talking straight to the person who can fix it.

No feedback is too small. "This wording confused me" is gold.

Beta: known limitations (what's still rough)

Honest list, so the rough edges read as known, not broken:

• It's an anticipation engine, not an oracle. Expect strong coverage of the territory, not the exact questions. Treat every score as a study aid.
• Fields vary. It's strongest in technical/STEM and social-science theses; humanities (philosophy, literature) is harder — exact-hit is lower there by nature. Still useful for breadth, but judge it with that in mind.
• Examiner profiles lean on public info. For early-career, industry, or low-footprint examiners, the profile is thinner and flagged low confidence — verify it against your own read.
• Revised your thesis mid-prep? For this beta, please finalize your research questions before you start prepping — re-uploading a changed thesis doesn't yet carry your prep forward cleanly.
• In beta you'll see a couple of things still settling: account/login, the in-app feedback widgets, and the usage-limit banner are new — if any misbehaves, that's exactly a bug worth reporting.

Your mission — by tester type

Everyone uses the same tool; we're asking each group to focus on a different question. Find yours:

🎓 Recently defended (last 1–3 years) — "did it call your defense?"
Upload your thesis (or paste your RQs + committee). Go through the anticipated territory cell by cell
and tell us, per cell: did your real examiner actually probe this? Where did it match, where did it
miss, and — if you have notes/transcript from your defense — what did examiners ask that it didn't
anticipate? The one anticipation that would've helped you most is exactly what we want to hear.

📝 Currently preparing (3–12 months out) — "does it actually help?"
Use it for your real prep over a couple of weeks — at least 3 sessions. Track which cells you flag as
weak and whether your advisor / mock committee end up probing those areas; whether cells you marked
Ready held up. After your defense, we'll ask: what helped, what was missing, what surprised you.
(Your prep data is private — your advisor never sees it.)

👩‍🏫 Faculty / examiners — "is this realistic from inside the field?"
You can skip the candidate flow. Review a sample of anticipated-territory cells in your field and tell
us, per cell: would a real examiner with these specialties actually probe this? What's missing? What's
wrongly anticipated? Optionally, review your own generated profile and tell us how close it is. We're
after expert calibration, not a candidate experience.

FAQ

Do I lose my work if I close the tab? No — everything saves in your browser automatically.

Are these the exact questions I'll be asked? No. They're well-grounded anticipations from your
examiners' public work and your thesis content — use them to prepare broadly, not to memorize.

What's the difference between a cell drill and a mock viva? A drill is focused practice on
one cell, with coaching after each answer. A mock viva is the full, realistic run — your whole
committee, end to end, scored at the finish.

Why didn't my mock viva score after each question? The mock viva runs realistic by default
(no mid-run scores — like the real room). You get the full read at the end. Switch to coached at
setup if you want per-question scores.

What do the cell colors mean? Solid = you can defend it · Wobbly = getting there · Weak = needs
work · Not yet = untouched. They reflect your most recent practice, weighted toward recent attempts.

Can I practice answering aloud? Yes — choose "Say it (aloud)" and tap Speak. Great for the real
oral experience.

Where do my saved questions go? The ★ Saved questions shelf on Home, and back into their cell's
practice pool. Build a custom viva from them anytime.

How do I sign in? With a one-time link or code we email you — no password to create or
remember. Request a new one anytime from the sign-in screen.

Is my thesis private? Can my advisor see my prep? Yes, it's private. Your work is encrypted at
rest with per-user access controls — other testers and your advisor can't see it. Your prep is yours.

Is my thesis used to train AI? No. Thesis text is processed to generate questions and score
answers under terms that prohibit training use — it's never used to train AI models.

How do I delete my data? Any time — in the app (account settings) or by email. It removes your
account and all associated data.

Why did the AI pause / what's this usage banner? There's a weekly usage allowance (AI calls cost
real money). You'll get a heads-up banner as you near it; if you hit the ceiling, AI features pause
and reset weekly — your saved work is safe. Bumped into it during honest prep? Tell us and we can
raise it.

Speak / the mic isn't working. Make sure you allowed microphone permission (check the
address-bar site settings), and try Chrome. Your speech is transcribed to text and scored on that;
audio isn't kept in this beta.

It scored me harshly or got something wrong — what do I do? Push back: use 👎 on the score
(with a comment) or the Feedback button. Disagreements are some of our most useful data — that's
how scoring gets calibrated.

Will I lose my work between devices? No — in the beta your prep is saved to your account, so you
can pick up on another device after signing in.

Questions or something not working as described? That's useful — tell us; the rough edges are
exactly what beta is for.