LOUDLY · MANTA 1 · VEGA · STYLE + LYRICS

Loudly Prompt Guide

Loudly is two products wearing one logo. MANTA 1 writes full songs with vocals from a style field and lyrics; the VEGA models build instrumental beds from genre, mood and length. Pick the path before you write a word, because the fields differ.

How Loudly takes direction

The song path takes a style description of up to about 1000 characters, lyrics of up to about 3000, a vocal gender and an instrumental switch. The bed path takes genre, mood, tempo and length. Both live on the same platform, and the instrumental switch defaults to on — so if you want a vocal you have to ask for one rather than assume it.

MODEL: MANTA 1 (song) or VEGA (bed) → STYLE: ≤1000 chars → LYRICS: ≤3000 chars → VOCAL GENDER: male / female → INSTRUMENTAL: on/off → LENGTH

Layer by layer

Loudly setups you can paste

One song on MANTA 1, two beds on the VEGA path. Paste the block for the model you picked.

Indie pop song · 108 BPM · vocals

MODEL: MANTA 1 STYLE: Bright indie pop at 108 BPM in E major — jangly electric guitar, warm analog drums, a female lead vocal with an airy delivery. Close-miked, slightly rough, short intro. LYRICS (section tags included): [Verse 1] (write your verse here) [Pre-Chorus] [Chorus] (biggest hook — repeat the title phrase) VOCAL GENDER: female INSTRUMENTAL: off · LENGTH: 150s

Product demo bed · 100 BPM · 90s

MODEL: VEGA 2 (instrumental) STYLE: Underscore for a software product demo — calm and modern, steady energy with no build, leaving the midrange free for a voiceover. Ends on a resolved chord. GENRE: corporate / electronic MOOD: calm, optimistic TEMPO: 100 BPM INSTRUMENTAL: on · LENGTH: 90s

Podcast intro · 88 BPM · 30s

MODEL: VEGA 1 (instrumental, low latency) STYLE: A short podcast intro that establishes the show and settles quickly, warm and conversational, no aggressive percussion. GENRE: lo-fi / soul MOOD: warm, welcoming TEMPO: 88 BPM INSTRUMENTAL: on · LENGTH: 30s

More ready-made prompts across moods and engines: AI music prompt examples.

What to leave out

The rule underneath all of them: describe the music, never the record. Reference an artist or a track and you inherit someone else's rights problem along with a vaguer brief.

Build a Loudly brief from a reference track

The hard part is the honest numbers: tempo, key, brightness, dynamics, and the instruments actually carrying the groove. Music DNA Studio measures them from any MP3, WAV, FLAC or MP4 — locally in your browser, nothing uploaded — then outputs a Loudly-shaped prompt alongside versions for the other fourteen engines. You make the creative calls; the spec underneath is measured, not guessed.

Both paths start from the same measured facts — tempo, brightness, dynamics, instrumentation — which is what makes switching between a song and a bed cheap.

Guides for the other engines

Eleven Music · Suno · Udio · Mureka · MiniMax Music · YuE2 · Soundverse · Treblo · Stable Audio · Soundraw · Mubert · Beatoven · AIVA · Loudly · MusicGen

Build The Loudly Brief

Drop in a reference, pick Loudly, copy the style block for MANTA 1 or the settings for a VEGA bed. Free, no signup, nothing leaves your device.

ANALYSE YOUR AUDIO — FREE →

FAQ

Does Loudly generate vocals? ▾

Yes, on MANTA 1 — its song model produces full tracks with lyrics and a chosen vocal gender. The VEGA models are instrumental, and the instrumental switch defaults to on, so a sung track has to be asked for explicitly.

Which Loudly model should I use? ▾

MANTA 1 for a song with vocals; VEGA 2 for a production-quality bed and VEGA 3 when you want the arrangement decisions handled in-generation. The model choice decides which fields you fill, so make it first.

How much can I write in each field? ▾

Roughly 1000 characters of style and 3000 of lyrics. The style field is a description of the sound, not a brief — genre, tempo, instruments and vocal character fit comfortably in two sentences.

Can I use Loudly tracks commercially? ▾

Paid plans carry commercial rights; the free tier is a demo tier with short previews, one download a day and non-commercial terms. Check the licence attached to your plan before delivery, and note that Loudly keeps the copyright on generated tracks.

Why does my bed fight the voiceover? ▾

Density and midrange. Music that fills 200 Hz–2 kHz competes directly with speech; ask for sparse midrange, one lead instrument and no vocal-like textures.