Soundverse fall 2026 launches: every new tool, model and app

Quick answer: From September to early November 2026, Soundverse shipped what it calls the biggest update in its history, for independent artists and creators: Agent One with three modes, AI Studio in the browser, Soundverse v9 and new outside models, a video suite, voice and instrument training, edit tools, Soundverse in Claude and ChatGPT, an API, Move from Suno and an Android app. Free covers Agent One Lite and a starter AI Studio, Creator ($12.49 a month) unlocks video, remix and voice training, and Soundverse v9 needs Pro ($31.24).

The brand-new Soundverse: everything that's new, in under two minutes.

Every launch at a glance

Dates are when each launch video went up on Soundverse's YouTube channel. For the longer story, from the 2023 assistant to Agent One, see Soundverse launches, 2023 to 2026.

Date (2026)	Launch	Deep dive
8–10 Sep	Tabla and dholak in Voice to Instrument	Training on you
10 Sep	Collaborative AI Studio and the plugin rack	AI Studio guide
10 Sep	Soundverse MCP for Claude	MCP guide
28 Sep	Mix One and Mastering One	AI Studio guide
29 Sep – 7 Oct	AI Video Describer, Video to Audio, video models, Video to SFX, Lip Sync	Song to video
9–14 Oct	AI Remix, AI Stem Generator, Audio to MIDI	Edit tools
12–15 Oct	AI Reels with Agent One, new AI Music Video Generator modes	Song to video
15 Oct	Move from Suno	Move from Suno guide
17–20 Oct	Agent One's three modes, Agent One inside AI Studio	Agent One
21 Oct	Computer Use and the Chrome extension	Chrome extension
22 Oct	Google Lyria 3.5, ElevenMusic v2.5, new video models	Model line-up
22 Oct	Audio to Prompt	Analyze tools
23–24 Oct	Voice Training, one-take Voice Clone, Artist DNA and instrument training	Training on you
24 Oct	Soundverse v9 vs Suno v6 blind test	Model line-up
24 Oct	Soundverse for Android	Android guide
2 Nov	Soundverse in ChatGPT	MCP guide
3 Nov	Soundverse API, Audio to JS, AI Studio text lane	API guide

The features page groups every tool by what it does, starting with Create.

Agent One: three modes, and now your browser

Agent One is Soundverse's AI music agent. Describe an outcome and it plans the steps, picks the tools and runs them. It now has three modes, switched from the chat box: Lite for quick ideas, Standard for everyday songs and edits, and Pro for multi-step jobs such as a song with a music video. Free runs Lite, with plans of up to 4 steps. Paid plans add Standard and Pro, with up to 12 steps on Creator and 120 on Pro and Max.

Computer Use (21 October) lets Agent One work on the real web, in Soundverse's cloud browser or in your own Chrome through the Soundverse extension. In the demo it worked through the release to-do list for "Midnight Neon Drive", a 120 BPM synthwave single, and uploaded its visualizer to YouTube as Private. Computer Use is on Free too.

Agent One makes a song, remixes it and turns it into a music video.

The Soundverse extension lets Agent One work in your own Chrome.

AI Studio: a full DAW in your browser

AI Studio is a multitrack studio in your browser, with nothing to install. In the 1 November tutorial, producer Mayur Jumani makes a track with the agent, records vocals, splits stems, plays Synth One from a MIDI controller and makes a music video in the Visual Lane. Mix One and Mastering One mix and master by genre, and every decision stays a setting you can adjust.

A text lane arrived on 3 November. In the "Neon Nights" example, "This city never sleeps. Neither do we. Tonight, we move." lands on screen with the vocal and is burned into the export. Free includes 3 projects of 3 tracks and one video lane. Deep dive: the AI Studio guide.

Soundverse v9 and the new models

Soundverse v9 is the newest Soundverse version, on Pro and Max. Creator adds Soundverse v7 and v8, Google Lyria 3.5 and ElevenMusic v2.5, and Free runs Soundverse v6. Lyria 3.5 works from text alone, while ElevenMusic v2.5 sings your lyrics as written at a length you set. AI Video Generator offers Veo 3.1, Seedance 2.0 and 2.5, and Gemini Omni Flash 1.1, on Creator and up. Deep dive: the model line-up by plan.

From song to video

Since 15 October, AI Music Video Generator turns one prompt into a full song and music video that opens in AI Studio with every section editable. Or give it a song, a character image and a prompt, and it plans the scenes, lip-syncs the performance and cuts to the music. Lip Sync, AI Reels, Video to SFX, Video to Audio and AI Video Describer complete the set. The video tools need Creator, with 1080p export there and 4K on Pro. Deep dive: from song to video.

Make AI sound like you

• One-take Voice Clone: record a single take, ask Agent One to clone it, then sing new songs in it with "Use custom voice".
• Voice Training: a deeper voice model for Voice Swap, trained on 10+ minutes of clean, dry vocals.
• Artist DNA: ask "Train my artist DNA on these 4 songs". Agent One asks you to confirm the rights first and emails you when the model is ready.
• Instrument models: train on an instrument you play, then hum a line and hear it in that sound through Voice to Instrument.

Voice Clone and voice, instrument and DNA training are on Creator. Voice Swap, Voice to Instrument and making music with DNA models are on Free. Deep dive: training on you.

The edit tools

On 2 November Soundverse reworked "Vande Mataram", Bankim Chandra Chattopadhyay's 1875 public-domain hymn. Similar Song Generator on v9 made a new song, "Radiant Motherland", in about 90 seconds, then AI Remix kept the melody and rebuilt it as an orchestral anthem and an Indian classical fusion. AI Stem Generator adds a new stem, such as an instrumental for your a cappella, and Audio to MIDI writes one .mid file per instrument.

AI Remix, Similar Song, AI Stem Generator and Inpainting are on Creator (v9 versions need Pro). Audio to MIDI, Extend and 2- and 6-stem separation are on Free. Deep dive: the edit tools.

Soundverse in Claude and ChatGPT

The Soundverse MCP connector lets you make music from a chat. Add it to Claude as a custom connector. A 2 November tutorial shows the same setup in ChatGPT, as a custom MCP plugin with the same server URL. In Claude, each tool can need approval (the default), be always allowed or be blocked, and every song lands in your Soundverse Library. Deep dive: the MCP guide.

The Soundverse API

The 3 November walkthrough goes from sign-up to a finished song in three calls. POST /v7/generate/song returns a job ID straight away, you poll until the status reads "completed" (71 seconds in that run), then download a 3:03 FLAC through a signed link valid for an hour. The same calls fit in about 20 lines of Python. Docs and keys are at platform.soundverse.ai. Deep dive: your first AI song in 3 API calls.

Move from Suno

Download your own Suno library and drop the files, a folder or the ZIP into Move from Suno. Soundverse sorts songs, stems, lyrics, prompts, covers, metadata and versions into projects you can open in AI Studio or continue with Agent One. It never connects to your Suno account or asks for your Suno password. Deep dive: the Move from Suno guide.

Soundverse for Android

In the 24 October walkthrough, one prompt becomes an instrumental, a title and cover art in 1m 14s. The app on Google Play has an Explore feed with Staff Picks, playback controls in the notification shade, and sharing as a link, a lyric video or an MP3. On iPhone, look for "Soundverse - AI Song Generator" on the App Store. Your Library is the same on every device. Deep dive: the Android guide.

Audio to JS and the analyze tools

Audio to JS (3 November) measures a sound's length, peak, pitch, note and noise, writes a recipe for it and ships it as a JavaScript file. In the demo, a one-second "Digital Chime" became chime-beep.js, a synth blip at 554 Hz (C#5) that matched on the first attempt. Audio to Prompt turns a song into a reusable style prompt and instrumental prompt. The analysis tools, Audio to Prompt among them, and the Copyright Checker are on Free. Deep dive: the analyze tools.

DJ One

DJ One is a two-deck booth in your browser that plays the songs in your Library, with SYNC, hot cues, loops, keylock, performance effects and Automix for hands-free transitions. Agent One sits in the booth to make a fresh track on request.

DJ One: two decks, SYNC and Automix, in your browser.

What's still coming

Distribution to Spotify, Apple Music and 150+ stores is coming to Max. Trace, shown in the 21 October video, and the Watermark Checker are in preview, not yet open to customers.

New to the app? The complete tour, from the sidebar to your Library.

FAQ

Which of these launches can I use for free?

Free covers Agent One Lite, AI Studio (3 projects of 3 tracks), three full-song generations on Soundverse v6, Extend, Voice Swap, Voice to Instrument, 2- and 6-stem separation, Audio to MIDI, the Copyright Checker, the analysis tools and Computer Use, all for personal, non-commercial use.

Can I release what I make on Spotify?

It depends on your licence. Creator has a "Royalty Free license — monetize your content", Pro a "Sample license — flip & release derivatives", and Max a "Distribution license — collect streaming revenue", with distribution itself coming to Max. Check the Terms before you release. This isn't legal advice.

Where should I start?

Describe the song you want to Agent One, then open it in AI Studio for hands-on control. Soundverse has 100+ tools and workflows, grouped on the features page.

Try it: start a chat with Agent One, or compare plans on Pricing.