Kirexa: Vocal Remover Offline

Skytrix Labs

In-app purchases

About this app

Remove vocals, split songs into clean stems, and strip out background noise, all on your phone and fully offline. Kirexa runs studio-grade AI audio tools 100% on-device, so your audio files are never uploaded to any server. Other apps send your audio to the cloud. Kirexa doesn't.

🎤 Vocal remover & karaoke maker
Pull a clean acapella or instrumental from any song. Remove vocals to make karaoke, sing over the backing track, or grab a vocals-only stem for remixing. An offline vocal remover with no cloud round-trip.

🎚️ Stem splitter
Separate any track into vocals, drums, bass, and other stems. Built for remixes, mashups, sampling, and practice. Multi-track source separation, right on your device.

🔇 AI noise removal
Strip background noise, hiss, hum, room tone, and chatter from voice memos, interviews, podcasts, and field recordings. One tap, no fiddly dB knobs. Kirexa's AI noise remover tunes itself.

✨ Voice Polish
Make any voice note or recording sound studio clean in one tap. Pick a sound and Kirexa balances your voice with EQ, compression, and even loudness. Podcast for a balanced, broadcast-ready voice, Warm for a fuller low end, or Crisp for bright and present. No sliders, no guesswork.

🎙️ Voice recorder
Capture a voice note right in the app. Tap Record and your take drops straight into the editor, ready to clean up, polish, or transcribe. No file juggling.

🎛️ Audio editor
More than a vocal remover: trim clips, remove silent gaps automatically, normalize loudness (LUFS or peak), and adjust speed. Export at the exact level Spotify, YouTube, podcasts, or broadcast expect, so your audio never gets turned up or down after you upload. A full lightweight audio editor in one app.

📝 Transcribe, caption & translate
On-device speech-to-text in 90+ languages. Turn voice memos, interviews, and lectures into text. Export subtitles as SRT, VTT, or LRC for video and social, or plain text for notes. Speak another language? Translate it into an English transcript. Your recordings never leave your phone.

📦 Batch export
Queue a whole folder and let Kirexa process it in the background. Clean, split, or remove vocals from many files at once.

🔒 Private by design
Every AI model runs on-device. Your audio is processed on your phone and stays there. No account required to start, no audio uploads, no servers. Only anonymous, opt-out usage stats, and your audio is never part of them. Transcription, recording, and cleanup work out of the box; stem separation downloads its model once, then runs offline.

Why Kirexa
• Offline vocal remover, extract instrumentals & acapellas
• Karaoke maker from any song
• Stem splitter for vocals, drums, bass & more
• AI background noise & hum removal
• In-app voice recorder
• Trim, silence removal, loudness & speed tools
• On-device speech-to-text (90+ languages)
• Caption & subtitle export (SRT, VTT, LRC, TXT)
• Translate speech to an English transcript
• 100% on-device, works offline, no audio uploaded
• Batch processing for big jobs
• Export in standard audio formats (WAV & more)
• One-tap Voice Polish with Podcast, Warm & Crisp presets
• Loudness targets for Spotify, YouTube, podcast & broadcast

Free, no paywall
Every tool is free, with no monthly minute caps and nothing locked:
• Full vocal removal & stem isolation
• Unlimited audio exports
• Batch export
• AI noise removal, transcription & translation
• Voice recorder, Voice Polish & loudness presets
If Kirexa is useful to you, you can leave an optional tip, a thank you rather than a purchase. Nothing in the app changes, and it stays free either way.

Built for musicians, producers, singers, podcasters, and video editors, anyone who needs a private, offline vocal remover, stem splitter, or noise remover without sending files to the cloud.

Download Kirexa, remove vocals, split stems, record, and clean up audio, 100% offline.

What’s new

Fixes crashes during silence removal and AI denoising, especially on longer recordings and lower-memory phones.

Stem separation now avoids incompatible graphics drivers and falls back safely when needed.

Background exports and processing notifications now finish reliably, and app startup is smoother.