Create music inspired by you
Turn a hum, a photo, or a few words into a full, studio-quality song that sounds like you.
NabadAi
Create Share Connect
The social platform for AI music creators.
Already have an account?
Choose your music styles
Pick at least 3 genres — choose as many as you like. We show your top 3 on your profile.
You're in
Let's make your first song
Now you have
—
credits
Language
Dialect
What's it about?
Sounds
Imagine your sound
Inspired by
Advanced
Tempo
120 BPM
Tap the Create button when your prompt is ready.
Create
Remixing
Track
Tap ✦ Generate to write a reply, or type your angle first.
Original lyrics
Singing voice
Persona
Used on your next song. Change or clear anytime.
Who leads?
Your voice vs. style tags
How adventurous?
Safe vs. experimental
Voice strength
Audio influence — how much you sound like you
Template loaded
Edit below, then tap Generate.
Melody guide
Singer
Vocal character
For Nabad Clip · Lyria
Advanced options
Voice
Type
Vocal style
Range
Length
Song length optional target
Avoid
Studio & More
Everything stays here so Home stays clean on mobile.
Home
Activity
Blend two tracks
Mashup
Pick two songs from your library or Discover — we blend them into one new track.
Playlist
Playlist
Top This Week
Ranked by plays and reactions in the last 7 days.
Hey, there
What are we creating today?
Quick Starts
Creative sparks
Sparks
Today’s spark
0 joined
Hook Rush
Make the first 15 seconds impossible to skip — no long intro, just the part people replay.
15 sec hook
Fast idea
Quick starts
Ideas to try now
Song templates
Occasion & dedication
Pick occasion and sound — no name needed.
Occasion
Sound
Language
Voice Lab
Voice scan on this device — for practice, not diagnosis.
Quiet room. One slow low → high glide.
Range labels
Tips for a clean scan
- Warm up lightly ~30s — no shouting.
- Glide slowly chest → head on “ah” (~6–8s).
- Hold your highest comfortable note ~2s.
Pick Men’s or Women’s so the printed range matches how you sing.
Your recent takes
Saved on this device only — no audio is kept.
This pass
Snapshot
—
—
—
Range
—
—
Key / maqam
—
—
Tone
—
—
Quality
—
—
Your strengths
From this take
Pitch motion
—
Brightness
—
Intonation
—
Detected maqam (heuristic)
Experimental: ranked on-device from your sung phrase using real cent intervals (quarter tones included). It's a rough estimate — short or ornamented phrases can read differently, so treat it as a hint, not a ruling.
Maqam detection is based on the recorded vocal phrase and may vary if the phrase contains modulation or ornamentation.
—
Genres that may fit
Intonation
—
—
Key ideas
—
—
Focus next time
—
—
Educational estimate only — maqam detection uses a coarse 12-semitone map. Range tags compare pitch to teaching lists you choose. “Timbre” here is brightness, not formants. Voice labels are not a fach exam.
No track loaded
0:00
0:00
Edit thumbnail
Adjust how the square thumbnail looks in feeds and lists. The full portrait cover stays for the player and reel.
Square preview
New cover
Optional artwork hint. Leave blank and Nabad will choose from your song.
@user
Become a fan
Verified Nabad Creator
0
Posts
0
Plays
0
Fans
Your signature
Create your persona now
Tap to set up your voice
No voices saved yet. Tap Record — voices sync to your account and return on a new phone.
Drafts aren’t permanent — they may stop playing · Publish or Download to keep
Popular this week
Credits
Redeem
How credits work
- Full song: 10–50 credits — depends on song type and options.
- Sounds: 2.5 credits · Lyrics are free.
- Failed generations are refunded automatically.
- Promo codes: once per account. Paid and promo credits can be gifted; gift credits are create-only.
Transaction history
No activity yet.
Step 3: Vocal Room (Multitrack)
Load stems as separate tracks, record vocal takes as a track, align start offsets, then export the full session.
Session
Tip: if tracks feel slightly late/early, use the offset control per track (seconds).
Vocal track
Recording creates a clip on the Vocal track (start offset adjustable like other tracks).
Step 2: Stems & Studio
After you generate a song, get instrumental version or multi-stems, then load stems into the studio and export your own mix.
Studio (basic mixer)
Play happens in-browser (no export needed). Adjust volumes while it plays.
Stems result
Advanced: local generator + hum → melody
1) Generate arrangement (local)
Tip: Leave “Server AI (paid)” off to use the free local generator. Turn it on only if you added a paid AI key
in Vercel.
Arrangement JSON
2) Hum → Melody (MVP)
Record up to 60s. We’ll detect pitch and turn it into a melody line you can use for arrangement.
Detected melody (JSON)
3) Render real audio
Ready.
What “no MIDI” means here
We generate an arrangement internally (notes + rhythm), but the output is rendered into an actual audio buffer and exported as WAV. Nothing outputs a MIDI file.
4) Voice AI (legacy)
Legacy endpoint for speech/TTS. If you use AI vocals, you likely don’t need this.
Edit Profile
Identity
Social links
AI
Settings
Manage your account, creator tools and preferences.
Account
Music presence
Appearance
Theme
Dark mode
NabadAI Orb
The orb shares tips and tracks your songs and instrumentals.
Orb stays visible on main tabs
Credits & plan
Creator tools
Your voices
Voices are saved to your account and sync across your devices. A recorded voice fades over time — re-record it about once a week so songs keep sounding like you.
Safety
Support
About
NabadAi Music
Version 1.0.0
Danger zone
Build detecting...