What's New

Launched PlayHT 3.0 model — 40% more expressive, 30% lower latency than 2.0

AI Music Generator — generate royalty-free music from text prompts

Sound Effects Generator — hyper-realistic SFX up to 22 seconds

Voice Cloning now requires only 30 seconds of audio (down from 2 minutes)

Dubbing now supports 30+ target languages (up from 22)

Fixed a bug where cloned voice stability would degrade on long-form text

Fixed streaming API dropping last syllable on certain short texts

WAV download occasionally produced a corrupted header on Firefox

Voice library search now fuzzy-matches accent and region descriptions

Voice Isolator — separate vocals, music, bass, and drums from any audio file

Dialogue Studio — multi-speaker script with per-character voice assignment

API Keys management panel with usage stats and per-key rate limiting

Generation history now stored for 90 days (up from 30)

Dashboard redesigned with weekly credit usage chart

Notification bell count was not clearing on page visit

Speech-to-Text sometimes assigned all audio to Speaker 1 when diarization was enabled

Billing page top-up modal was not closing after a successful purchase

Mobile nav now includes all 8 feature links in the dropdown

Complete app redesign — dark glass morphism UI across all 17 pages

Auto Dubbing with speaker diarization and lip sync support

Voice Changer with real-time neural conversion and pitch control

Speech to Text with word-level timestamps and SRT/VTT export

Credit-based billing with top-up packs and usage breakdowns

TTS generation speed improved 2× with new streaming architecture

900+ library voices now with preview on hover

Email verification link was expiring after 30 minutes instead of 60

Voice clone delete button was not removing the model from the TTS voice picker

Increased max text length per TTS generation from 30,000 to 50,000 characters