Create music with the most advanced AI models

Shape full songs from a short brief — not stock loops. Lock the vibe once, generate in your workspace, and walk away with tracks ready to use.

Generate finished professional tracks in a minute with Lyria 3 Pro

Lyria 3 Pro is Google DeepMind’s flagship music model — built for complete songs, not short loops. It plans intro, verse, chorus, and bridge so the track feels arranged, not stitched together.

Describe the song in text, or guide mood and style with a reference image. You get high-fidelity stereo audio up to about three minutes, with vocals and timed lyrics — or a clean instrumental when you need a soundbed.

In Veontra you set genre, mood, tempo, and duration, then download MP3 (and optional cover) from the same workspace. Every Lyria output is SynthID-watermarked for AI provenance.

Steer vocal songs with your reference track using MiniMax Music 2.0

MiniMax Music 2.0 builds full vocal tracks guided by an audio reference you upload — great when you already have a vibe, groove, or performance you want to echo.

Upload a reference song, describe the style you want, and choose lyrics mode: instrumental, auto, or your own words. The model follows the reference while shaping a fresh arrangement.

In Veontra the reference upload lives in the same generator as genre, mood, and vocals. Length follows the lyrics, not a separate timer — then preview and download the MP3 from your workspace.

Strong vocals from your lyrics with MiniMax Music 2.6

MiniMax Music 2.6 is built for complete tracks with singing, backing music, and detailed arrangements from a style brief and lyrics.

Write a short style prompt, then use auto lyrics or paste your own with structure tags like verse and chorus. Prefer a soundbed? Switch to instrumental and skip vocals entirely.

In Veontra you still set genre, mood, vocals, and tempo in the generator. Track length follows the lyrics — longer text usually means a longer song — then download from the same workspace.

Go long with Sonilo v1.1 — tracks up to 10 minutes

Sonilo v1.1 is made for extended instrumentals when a short loop is not enough — beds, scores, and long-form grooves up to about ten minutes.

Describe the style, pick genre and mood, set the duration you need, and generate a continuous instrumental without vocals or lyrics modes.

In Veontra duration is billed per second, so short tests stay cheap and long tracks scale when you need them. Preview in the feed, then download the MP3 from your workspace.

How it works

From idea to downloadable track

Three clear steps — choose how you generate, describe the song, then take the files you need.

Choose a model for generation

Pick the AI music model that fits your track — Lyria 3 Pro, MiniMax Music 2.0, MiniMax Music 2.6, or Sonilo v1.1 — all in the same selector.

Describe the music

Write a prompt and set genre, mood, vocals, tempo, and duration (1–3 minutes). Choose lyrics mode — instrumental, auto, or your own text — optionally attach a reference image (image → MP3), and request an album cover.

Generate lyrics too

Download or export

Preview in the dashboard, then download the MP3 alone, grab a ZIP with cover when you generated one, or export either option to Google Drive.

Any style you can imagine

From pop anthems to lo-fi and cinematic scores — pick a vibe, listen to a sample, then generate your own track in Veontra.

One AI workspace for your team

Extract documents, transcribe audio, and generate images from the same dashboard — with human review before anything is exported.

Save your time with AI

Start your 14-day trial with up to 100 credits.