Wicara Coming soon iPhone

Answer out loud. Under pressure. Get numbers back.

An AI interviewer asks questions for your role and follows up when you're vague. Each question reaches you through background noise and a rough line, so you have to listen hard. Every answer is scored on what you said and how you said it.

Wicara Practice heat · iPhone

One email when it's out. Nothing else.

See how scoring works

In development · iPhone, iOS 18+ · English interviews

Pace

172wpm

target 120–160

80

Placeholder screen. The app's design isn't final.

Practising at your desk skips the hard part.

  • Silent rooms lie.

    Practising alone at a desk doesn't prepare you for a café call, a bad video link or an open office.

  • Lists don't follow up.

    A question bank never asks "can you give a concrete example?"

  • "You did great" isn't feedback.

    You can't improve a vibe; you can improve 7 fillers a minute.

WicaraIndonesian: to speak.

Speaking is a skill, not something you know. Breath, pace, the pause before your first word: none of it improves from rereading your notes. It improves the way any skill does. Out loud, again, with someone keeping count.

How a session runs

  1. Tell it the job.

    Role, seniority, your skills and weak spots, and the interview type: technical, behavioral or mixed.

  2. Small talk first.

    The interviewer opens with a casual question. It's not scored; it's how Wicara learns your normal voice.

  3. The interview.

    3–5 questions, spoken through the background noise you picked. Answer out loud, tap Done. Vague answers get one follow-up.

  4. The report.

    A score per answer and per session, every number explained, one tip per answer, a recording of yourself. Every session is kept, so you can see the trend.

Pick the room.

Questions arrive over a quiet office, a café, a street or a video call (that thin call-quality voice), at low, medium or high. The noise and distortion sit on the question; your answer is recorded clean.

Low
Medium
High
The questionCafé noiseAn illustration of the same question at three levels, not a recording.

  • Follow-ups that probe the gap.

    One per question, like a real interviewer.

  • No script on screen.

    You hear the question; you don't read it. Didn't catch it? Say "sorry, could you repeat that?", once.

  • A clock in the room.

    The delay before your first word is measured; answers cap at 3 minutes.

Every score shows the measurement behind it.

What you say wins interviews. How you say it is what you can't hear yourself.

Delivery five measured factors

FactorWhat it measuresTarget
Pacewords per minute while speaking120–160 wpm
Response delayseconds before your first word≤ 3 s
Fillers"um", "you know", hesitation sounds, per minute≤ 2 / min
Intonationhow much your pitch moves vs your warm-up≥ 90% of normal
Articulationhow clearly you're understood vs your warm-up≥ 95% of normal

Content three criteria, 1–5 each

Relevance
Did you answer what was asked.
Structure
A clear point, in order.
Substance
Specific, correct, with examples.

Every content score quotes your own words as evidence.

One answer, scored

Example, computed with Wicara's scoring rules

Delivery readings
DeliveryReadingTargetScore
Pace136 wpm120–160100
Delay2.4 s≤ 3 s100
Fillers7.1 / min≤ 249
Intonation72%≥ 90%55
Articulation93%≥ 95%92
Content scores
ContentEvidenceScore
Relevance"The launch I'd point to is…"4
Structure"…and also, going back to before…"3
Substance"we improved onboarding a lot"3

Delivery78.8

Content58.3

Answer 0.6 × 58.3 + 0.4 × 78.866.5

Tip. Fillers cost the most here: 7 a minute against a target of 2. When you feel an "um" coming, stop and stay silent. A silent pause isn't counted as a filler.

Fair and private

Fair to your voice

Wicara compares your intonation and clarity to your own warm-up, not to a standard voice. Accent and mic don't cost points; losing your voice under pressure does.

Pace, delay and fillers use the same targets for everyone, because listeners do.

Your voice stays on your phone

Speech is transcribed on the device and recordings stay there. Only the text goes to Claude (Anthropic), to write the questions and score the content.

No Wicara account and no sign-up. It's an iPhone app, so your Apple account is all it takes.

Questions

When is it out?
No date yet. Leave your email and you'll hear once.
Do I need headphones?
Recommended, not required.
How long is a session?
3–5 questions, plus follow-ups.
Which languages?
English for now.
Coding interviews?
No. Technical questions are spoken and conceptual; there's no code editor.
Is the score the same every time?
Delivery is measured, so it's repeatable. Content is judged by an AI against a fixed rubric with quoted evidence, so it's consistent but not identical across runs.
Does it give live feedback while I talk?
No. Like a real interview, you find out afterwards.
What does it cost?
Not decided yet.

Your next interview will be noisy. Practise like it.

Wicara Practice heat · iPhone

One email when it's out. Nothing else.