Platform announcement

Three AI experts
interview you.
Simultaneously.

InterviewOS is a desktop application that runs a real voice interview — three distinct AI personas, each with their own voice and specialization, share conversation history and probe your actual project experience.

by hiwarkhedeprasad · tauri 2 + react 19 + ollama

[ Alex ]

↓ macOS ↓ Windows GitHub → MIT · Free · Requires Ollama

The Panel

Alex Chen

Peer · Junior Dev · Fast Voice

Sarah Mitchell

Team Lead · Tech Lead · Mid Voice

James Rodriguez

Veteran · Principal · Deep Voice

By the numbers

AI Agents active3

Languages7

LeetCode questions0

Session modes4

PlatformDesktop

"Your projects,
not puzzles"

All questions are project-based and experience-focused. The agents read your resume and ask about what you actually built.

What's inside

Voice Interview Engine

Web Speech API runs in continuous mode with interim results. VAD detects when you start talking and interrupts the agent mid-sentence. 1.5s of silence triggers the AI response cycle.

Web Speech API

Monaco Code Workspace

VS Code's editor embedded in the app. Resizable panels. JavaScript runs sandboxed with console capture. Test runner shows pass/fail. Ollama reviews your code for quality and correctness.

Monaco Editor

Resume Parsing

Drop a PDF. pdfjs extracts skills across 6 categories, up to 5 projects with detected tech stacks, and estimates your seniority level. Every question is contextualized to you.

pdfjs-dist

Proctor Lockdown

Rust backend: maximized, no decorations, always-on-top. Windows uses WDA_EXCLUDEFROMCAPTURE. macOS uses panel-level window. The app is invisible to screen recorders and capture tools.

Rust · WinAPI

Scored Report

Per-question scores 1–10, candidate vs preferred answers, strength and weakness breakdown by topic, and a downloadable self-contained HTML file with print styles.

HTML Export

The Interview Panel

All three agents share conversation history — no question is ever repeated. A random agent responds, with 30% chance of a follow-up from a different one.

Peer Interviewer

Alex Chen

Junior Developer

Fast, higher-pitched voice. Friendly energy. Alex asks about how you work day-to-day, communicate in a team, and your fundamentals.

Focus Areas
Fundamentals · Communication · Day-to-day workflow

Team Lead

Sarah Mitchell

Technical Lead

Medium pace, professional tone. Sarah pushes on architectural decisions, forces you to justify trade-offs, and probes how you collaborate under pressure.

Focus Areas
Architecture · Trade-offs · System design

Veteran

James Rodriguez

Principal Engineer

Slow, deep, authoritative. James hunts for edge cases, production war stories, and whether you actually shipped anything hard at scale.

Focus Areas
Edge cases · Production experience · Scale

Interview Flow

Setup

Enter your name, drop your PDF résumé. Skills, projects, and experience level are extracted. Agents load your context before a single question is asked.

→

Video

Full-screen video call. Camera, mic, screen share. 20-minute countdown. All three agents are live. VAD handles turn-taking — no buttons to click.

→

Coding

Transition to Monaco workspace. Resizable panels: question sidebar, editor, output. Run code, see test results, get AI code review on demand.

→

Report

Rust backend generates a full report: per-question scores, preferred answers, topic-level strengths and weaknesses. Download as self-contained HTML.

Tech Stack

Frontend	React 19Vite 7
Desktop	Tauri 2Rust
Editor	Monaco
AI / LLM	Ollama Local or Cloud
Speech	Web Speech API + Synthesis
PDF	pdfjs-dist
Output	.msi · .dmg · .deb

Sample Report

9.1

System Design

Strong architectural reasoning. Justified the caching decision with concrete trade-offs.

6.4

Edge Cases & Production

Answer was adequate but lacked war stories. More detail needed on root cause analysis.

8.7

Communication

Explained complex concepts clearly. Good analogies. Would perform well cross-functionally.

Three AI expertsinterview you.Simultaneously.

Three AI experts
interview you.
Simultaneously.