Every tool you need to walk in ready.
PrepLoom is built around how real interviews are scored — substance, signal, and presence. Here's what you get in the practice room today.
Role-aware questions
Questions built around your target role.
Pull from a structured RAG-powered bank aligned to your target role. Each session mixes behavioural and technical prompts.
Structured, role-specific question sets.
PrepLoom pulls from a curated ChromaDB vector store, mixing 6–7 technical questions with 3 behavioural ones every session. No two sessions are identical.
- Retrieval-augmented generation (RAG) question fetch
- Technical + behavioural balance every session
- Difficulty and subtopic shown per question
- Ideal answers available on demand
Answer evaluation
LLM scoring that actually explains itself.
Submit spoken or typed answers for feedback on reasoning and accuracy. Ideal answers reveal a strong baseline to compare against.
Strengths, gaps, and what you missed.
Every answer gets a technical score, a list of strengths, areas to improve, and missing points — not just a number. The feedback is actionable on the same day.
- 0–100 technical score per answer
- Streengths and improvement points called out
- Missing key concepts surfaced explicitly
- Short qualitative feedback paragraph
Communication signals
Notice how you sound before the real room does.
Filler words, pacing, and structure — the signals a hiring manager notices without naming them.
NLP analysis on every answer.
PrepLoom counts fillers, calculates your words-per-minute, and scores clarity and structure — not just vocabulary. The goal is to make habits visible before they cost you.
- Filler word detection (um, uh, like, you know)
- Speaking pace in WPM with ideal range
- Clarity and structure score
- Whisper-powered transcription
Video & posture
Presence is a skill. Practice it.
Optional camera-based signals on openness, engagement, and posture — a nudge, not a verdict. Comfort still comes first.
Live body language via MediaPipe.
A WebSocket streams your webcam frames to a pose-estimation backend. You get real-time openness, engagement, posture, and fidgeting signals — updated every frame while you speak.
- MediaPipe pose estimation
- 4 live signals: openness, engagement, posture, fidgeting
- WebSocket continuous analysis (~15 FPS)
- Confidence score feeds into the final report
At a glance
PrepLoom vs the alternatives.
| Feature | PrepLoom | Generic Q&A sites | Mock interviews (human) |
|---|---|---|---|
| Role-specific questions | ✓ | ✓ | ✓ |
| Instant AI feedback | ✓ | ✗ | ✗ |
| Communication scoring | ✓ | ✗ | Sometimes |
| Body language analysis | ✓ | ✗ | Sometimes |
| Session report & tips | ✓ | ✗ | Sometimes |
| No scheduling needed | ✓ | ✓ | ✗ |
| Free to use | ✓ | Mostly | ✗ |
Open the practice room.
Select a role, answer questions at your own pace, and walk away with a full report in under 20 minutes.
Start practicing now