Polyfaced

Targeting late 2026 — join the waitlist

AI Music Coming Soon

Compose AI music from text prompts with vocal direction, genre control, and exportable stems for remixing. Built for short-form video creators, podcasters, and game soundtrack designers.

Get notified when AI Music launches

Be the first to try it. No spam — one email when it ships.

What we're building

AI Music is the third surface coming to Polyfaced after AI Avatar. Where AI Avatar is identity-stable image generation, AI Music is creator-grade text-to-song with full stem export — meaning you don't get a single locked stereo file, you get the vocal track, the drums, the bass, the synth pad, and the lead instrument as separate files ready for remixing in any DAW (Ableton, Logic, FL Studio, Reaper). The creator experience mirrors what's already shipping on /studio/video and /studio/image: prompt box, genre selector, vocal direction toggle (instrumental / with vocals / instrumental-friendly chord progression for your own vocals), duration setting, and an inline preview before download. Targeted at short-form video creators who need original tracks under 60 seconds, podcast producers who need scoring beds, and indie game devs who need a soundtrack without licensing fights.

Use cases we're building for

Short-form video soundtracks

Generate original music for TikTok, Reels, and Shorts where copyright-flagged tracks instant-disqualify your post. 60-second clips with the genre and vibe you describe — no licensing chase.

Podcast intro and outro stings

Generate signature intro and outro music for your show that no one else has. Refresh per season for a fresh feel without paying a composer for each pass.

Indie game soundtracks

For solo devs and small teams shipping on Steam, itch.io, or mobile stores — generate loopable background tracks and emotional cues without licensing AAA library tracks or hiring a composer.

Ad jingles and brand mnemonics

Iterate jingle ideas in minutes instead of weeks. Generate 20 variations on a 15-second hook, pick the one that wins your team review, then commission a composer to refine — or ship directly.

Roadmap detail

What's locked: stem export (full multi-track), genre selector, vocal direction control. What's still being evaluated: which underlying model wins (Suno v5, Udio, Meta MusicGen-X, an open-weight option), whether the v1 release supports vocal cloning from a reference voice, and whether 4-minute full songs ship at launch or just 60-second clips. Sign up for the waitlist to be notified when each milestone lands.

Meanwhile, try our shipped tools