Sound Studio · Voice + Music
A voice cast and a music studio — score your brand
Cast a voice from an over-100-actor panel and it reads your script word-for-word — or brief ✨ Author with Claude and approve the draft it writes. Compose exact-length jingles and full songs in ten genres. Everything auto-saves as a project, keepers file into the brand kit — and a saved jingle can score your next video, with the storyboard choreographed to the lyrics.
The jingle machine · Brief → song → soundtrack
Describe the jingle — Claude writes it, the studio sings it
Give it a one-line brief and a genre. ✨ Author with Claude drafts the lyrics — a hook that lands on your brand name — you edit the draft, approve, and a finished track lands in the feed at exactly the length you asked for.
The brief · pick a genre
“a jingle for our Suntastic Sundress — light and airy in the sun, spinning around and having fun”
The draft — yours to edit
[HOOK]
Sun-tastic days, we’re dancing in the daisies…
[VERSE]
Spinning all around, sunshine and glowing days…
ElevenLabs Music · 0:15 · MP3/WAV
Claude writes the song
A one-line brief comes back as a lyric sheet — a hook that lands on your brand name, verses that sell the story. ✨ Author is a flat ⚡ 10.
Edit before you spend
The draft is an editable card — change a line, tighten the hook, then approve. Nothing composes until you say so.
Exact lengths or full songs
Jingles land at exactly 15, 30, or 60 seconds — timed for ads. Ask for a full song and the lyrics are sung word-for-word. Instrumental when you just need the bed.
How it works, start to finish
Pick a brand and a style, cast the actor or pick the genre, then author or generate — clips land in a feed with a player, projects keep themselves, and keepers file into the kit.
Pick your brand
The Brand bar adopts your active kit — projects, saved clips, and your brand's own voices all key off it. Switch kits from the bar anytime.
Choose the style
Voiceover or Music — the bar reshapes around your pick. Voiceover adds the Actor bar; Music adds the Genre bar with lengths and an instrumental toggle.
The engine picks itself
Leave the Engine on Auto and the studio routes the job — the right voice engine for the actor you cast, the right music engine for the length you asked for. Or pick one yourself.
Cast the actor
Choose Actor opens the voice panel: an over-100-voice curated cast plus your brand's own voices — every one with a one-click audition clip.
Pick the genre
Ten genres seed the music direction — Jingle is the default (the video-ad workhorse). Your words steer the song; the genre sets the production style.
Author it — or drive it
Two lanes: type a finished script and Generate, or give ✨ Author with Claude a one-line brief and get an editable draft back — lyrics with a hook that lands on your brand name. You edit, then approve.
The feed
Clips land as cards with a player, the engine and actor named, and a download — MP3 or WAV. Fire the next idea while the last one composes.
Projects keep themselves
Your session auto-saves as a Sound Studio project the moment work starts, and keeps saving as you go. Reopen it anytime from the kit's Brand Video & Audio page.
Save keepers to the kit
One click files a clip in the brand kit's asset library under Audio — preview, download, and reuse it anywhere. Saved jingles can score your next video generation.
From the studio to the screen · Save → load → score
Save it to the brand kit — then it plays in your Cinema Ads
The studio isn’t a dead end — it’s the sound department for your video work. Save a jingle or a voiceover into the kit, then load it as the Audio reference on a video generation: the storyboard is choreographed to your track and the finished film plays it.
0:15 · saved with its lyrics
Brand kit · Audio
Cinema Ads · the references
“woman in a field of daisies, spinning, enjoying the sun”
One library, every video
Saved clips live under the kit's Audio tab — preview, download, and pick them from any video generation's reference panel.
Music or narration
A jingle becomes the soundtrack; a saved voiceover rides the ad the same way. Any clip in the kit can score the film.
Choreographed, not pasted on
With Cinema Ads the storyboard is written to your track — the action lands inside the lyric lines, and the finished film plays your audio.
The casting call · Audition → cast → read
An over-100-voice cast reads your script — audition anyone in one click
Open the voice panel and it’s a casting session: over 100 curated, named voices plus your brand’s own voices — each with a specimen clip. Cast one, hand over the script, and the read comes back word-for-word.
The voice panel · 100 curated + your brand’s
The script
warm · upbeat“Suntastic turns sunshine into style. <pause> The Summer Sundress — light, airy, made for spinning.”
Type it yourself — or let ✨ Author write it from a brief.
MiniMax Speech 2.8 HD · 0:22
Audition in one click
Every voice has a specimen clip — hear the actual read before you cast. Filter the panel by gender and style to shortlist fast.
Your brand's own voices
Voices saved to the brand kit join the same panel right next to the curated cast — one casting call, every source.
Direct the emotion
✨ Author scripts carry a performance direction — the read comes back happy, calm, urgent, or however the spot needs it. Pauses land where they're written.
Everything in the Sound Studio
The full cast, both music lanes, Claude authoring, auto-saved projects, brand-kit audio — and your jingle scoring the film.
Voice
An over-100-voice curated cast
Hand-picked, named voice actors — narrators, announcers, conversational reads — each with an audition clip. Your brand kit's own voices join the same panel.
Voice
Word-for-word reads
The script is the take — your words read exactly as written, with the emotion the author calls for. Powered by MiniMax Speech 2.8 HD and ElevenLabs.
Music
Jingles & full songs
Exact-length jingles (15, 30, or 60 seconds) through ElevenLabs Music, and full songs with sung lyrics through MiniMax Music 2.6. Instrumental toggle when you just need the bed.
Music
Ten genres, jingle-first
Jingle (the default), Pop, Acoustic, Electronic, Hip-Hop, Rock, Cinematic, Lo-fi, Jazz, Ambient — the genre seeds the production style, your brief steers the song.
Claude
✨ Author with Claude
A one-line brief becomes a finished script or a lyric sheet — hooks that land on your brand name, pause marks, an emotion direction. It comes back as an editable draft; nothing fires until you approve.
Keep
Auto-saved projects
Sessions save themselves as Sound Studio projects under the kit's Brand Video & Audio page — voiceover and music counts on the card, reopen exactly where you left off.
Library
Audio in the brand kit
Saved clips live in the asset library under their own Audio tab — artwork thumbnails, a preview player, and downloads. The studio also learns the words and their timing, so the AI can use them later.
The payoff
Score your videos
Attach a saved jingle as the soundtrack of a video generation (the newest video engines support it). The storyboard choreographs the action to your lyric lines — a spin lands on 'spinning all around'.
Simple
The price on the button
Every fire shows its token price right on the button before you commit — voiceovers priced by script length, music per piece, authoring a flat ⚡ 10.
Where this is heading
Today the Sound Studio is an in-app studio — including ✨ Author with Claude, which is live on the page. The conversation lane is next.
Drive it from a conversation
ComingAsk Claude for 'a 15-second jingle about our summer dress, sunny acoustic pop' and get the clip back — the studio's lanes, from chat.
More languages
ComingThe cast grows beyond English — hundreds of voices across the major markets, auditioned the same way.
Your voice, cloned
ComingBring a founder's or creator's voice to the cast so the brand can narrate as itself.
The curated cast is English-first for now — the language packs (hundreds of voices across markets) are on the roadmap above.
Exact lengths (15/30/60s) are the short-form lane — timed for ads. A full song composes to its natural length; your lyrics are sung word-for-word either way.
Reference audio rides the newest video engines (the Seedance 2 family) and carries up to 15 seconds of track — perfect for a jingle, not a full song. The button always shows the token price first.