Changelog

What is new

We ship every week. Track the development journey.

Feb 28, 2026

NEWv2.1.0Feb 28, 2026

Voice Mode 2.0

Real-time speech recognition integrated with Deepgram for ultra-low latency.
Filler word detection (um, uh, like) added to communication analytics score.
AI now responds with synthesized voice mimicking realistic interviewer cadences.
Added Speaking Pace analysis highlighting when you speak too fast under pressure.

Feb 14, 2026

MAJORv2.0.0Feb 14, 2026

Feb 1, 2026

IMPROVEMENTv1.9.0Feb 1, 2026

Added local tech giants: Systems Limited, Techlogix, 10Pearls, Arbisoft, Netsol, Creative Chaos.
Updated Google algorithmic complexity questions based on recent 2026 reports.
Amazon Behavioral pack now strictly enforces the STAR method in initial prompts.

Jan 20, 2026

UPDATEv1.8.0Jan 20, 2026

Moved entirely to the new 'Raycast' inspired dark theme aesthetic (#08080C backgrounds).
Unified all component border radii to 16px for consistency.
Replaced standard generic typography with DM Sans for UI and JetBrains Mono for code.
Optimized load times by switching to Next.js Turbopack.

Jan 8, 2026

SOCIALv1.7.0Jan 8, 2026

Dec 25, 2025

NEWv1.6.0Dec 25, 2025

A new randomized DSA question deployed globally at 12:00 AM PST.
Global streaks introduced. Missing a day resets multiplier.
Heatmap added to profile illustrating commit-like consistency for interview prep.

Dec 10, 2025

UPDATEv1.5.0Dec 10, 2025

Complete rewrite of the charting library transitioning to Framer Motion driven SVG metrics.
Added Skill Radar visualizing the balance between DSA, System Design, and Behavioral.
Introduced 'Weak Topic Detection' scanning historical session transcripts.

Nov 15, 2025

LAUNCHv1.0.0Nov 15, 2025

PrepWithAI is live! The first MVP shipped focusing purely on algorithmic coding interviews.
Integrated Groq LLaMA 3.3 70B for highly critical mock feedback generation.
Deployed on Vercel utilizing Edge Functions for sub-second latency.

Preparing your workspace…

Voice Mode 2.0

Real-time speech recognition integrated with Deepgram for ultra-low latency.

Filler word detection (um, uh, like) added to communication analytics score.

AI now responds with synthesized voice mimicking realistic interviewer cadences.

Added Speaking Pace analysis highlighting when you speak too fast under pressure.