feat: auto-detect CUDA GPU for whisper model to drastically improve speed on GPU hardware

fix: explicit garbage collection and model deletion to prevent OOM server restarts

fix: resolve UnboundLocalError when isolate_vocals is unchecked

fix: add real-time upload progress percentage and error handling for huge files

chore: move backend files into dedicated backend directory and clean up root

feat: add smooth progress bar simulation for AI model processing

feat: completely remove AI De-Reverb feature from frontend and backend

feat: remove Instrumental option from download customization UI

fix: resolve Smart Compare Studio playback and 'New Upload' reset link

feat: remove metadata CSV upload and forced alignment logic

feat: remove 'Your Tasks' panel from UI and backend API

fix: remove unused pyannote and basic_pitch imports

fix: restrict numpy<2.0 and remove health feature

fix: remove unused dependencies causing build conflict on HF

feat: completely remove Isolate Instrumental and 4-Stem Separation features

feat: completely remove Fast CPU mode toggle and fast MDX-Net separation logic

feat: completely remove Stem to MIDI feature from frontend and backend

feat: completely remove Separate Speakers (pyannote) feature from frontend and backend

fix: resolve NoneType object has no attribute to in pyannote speaker diarization by checking if pipeline is None

fix: resolve NoneType object has no attribute to in audio_separator by creating fresh instances and checking model_instance

fix: resolve audio-separator output paths to prevent FileNotFoundError on converted_(Bass)_htdemucs.wav

chore: trigger Vercel redeploy

Add python3-tk to Dockerfile to fix libtk8.6.so error

Massive architectural overhaul: Migrated separation to global MDX-Net cache via audio-separator, eliminating subprocess overhead and significantly boosting CPU speed. Added Fast Mode UI toggle with auto-fallback for large files.

Fix Demucs syntax error by removing --segment

Fix Hugging Face OOM crashes by forcing demucs single job, 10-sec segments and manual GC

Fix TypeScript path aliases for @/ components

Build Programmatic SEO Matrix with 40,000 keyword landing pages

Install react-markdown for blog rendering

Build CMS Blog Engine and 100-FAQ System

Add Google Search Console verification HTML file

Add Google Site Verification meta tag

Update branding to VocalBee

Hardcode sitemap baseUrl to vocalbee.vercel.app for canonical SEO

Update SEO files to use vocalbee.vercel.app as fallback URL

Update SEO files to dynamically use Vercel URL

Replace text logo with new SVG image logo in UI

Rebrand to VocalHive and add new favicon

Add robots.txt and sitemap.xml for SEO

Implement multi-task queue system with active tasks panel

Fix task persistence and session rehydration

fix: add root route to prevent hugging face 404 logs

remove: background 24-hour file cleanup timer

fix: make task expired error more user-friendly

feat: save extraction options to firebase and display active tools as badges in history

remove: faq and contact pages

remove: pricing page and billing completely

remove: completely remove resend and email_service

feat: trigger auto login modal on upload if not signed in

remove: dashboard page and links completely