What's New

Engine updates, new features and improvements to Suno Reverse. Each entry describes what changed for you as a user — not how the engine works internally.

Original Song Generation and Authorized Cover Controls

NEW FEATURESTUDIO

Suno Reverse now gives original song generation and reference-conditioned covers their own clear entry points. Creators can start from a prompt or their own lyrics, choose a vocal or instrumental arrangement, and direct supported style, tempo and duration controls.

Cover projects add a similarity control for an authorized reference. The page now states the boundary directly: similarity guides the result but does not promise exact notes, timing, stems or waveform copying, and it does not grant permission to use a recording, composition or performer identity.

Creation remains separate from post-production. A generated result can be downloaded, humanized when unwanted synthetic texture is present, or mastered after the composition and mix are accepted.

Faster Mastering Handoff and Clearer Delivery Status

IMPROVEMENTSTUDIO

The mastering workflow now moves from upload to paid processing with fewer avoidable handoff steps and keeps mastering credits separate from Humanizer credits.

Job polling, reconciliation and retry paths surface the WAV and MP3 as soon as the completed result is available. Processing time still varies with file duration, queue and service health, so Suno Reverse does not publish a fixed turnaround guarantee.

Successful mastering jobs provide a 48 kHz stereo PCM-24 WAV and a 320 kbps MP3. Mastering remains the final stage for an approved stereo mix, not a repair for wrong lyrics, arrangement or isolated balance.

KFK Neural Model 4B Humanizer

ENGINEIMPROVEMENT

KFK Neural Model 4B is now the public model name for the Suno Reverse humanization workflow. Private checkpoint, architecture and infrastructure identifiers remain internal.

The source-first goal is unchanged: target distracting synthetic texture while aiming to keep lyrics, melody, source vocals, dynamics, arrangement and stereo relationships connected to the uploaded complete mix.

Results depend on the source. Humanization does not rewrite provenance, create missing rights, guarantee a detector result or approve a release for any platform.

Authorized Suno Downloader With Honest Format Labels

NEW FEATURESTUDIO

The Downloader accepts a supported publicly retrievable Suno link or track ID and creates a 320 kbps MP3 or a 24-bit 48 kHz PCM WAV. It does not ask for Suno account credentials or unlock private projects.

WAV delivery is now labelled as decoded PCM rather than a lossless source master. The available progressive stream may already be compressed, and transcoding cannot recreate information that is absent from it.

The tool is for a creator’s own or otherwise authorized tracks. Suno Reverse is independent of Suno, and technical access to a public link is not proof of permission.

Remote Vocal and Instrumental Stem Splitter

NEW FEATURESTUDIO

A dedicated Stem Splitter now accepts one authorized audio file up to 100 MB and returns two estimated results: vocals and backing instrumental.

Creators can preview the source and separated files, download individual results, or collect the available files in a ZIP. Processing runs on a configured remote separation service rather than on the local device.

Separation is an estimate, not recovery of the original multitrack session. Dense mixes can leave vocal bleed, missing shared instruments or phasey ambience, so the workflow no longer describes results as broadcast quality, instant or guaranteed lossless stems.

Responsive Mobile Foundation and Expanded Playback

IMPROVEMENTSTUDIO

The web workspace now has responsive navigation, safe-area-aware layout behavior and an installable web-app manifest with direct shortcuts to core products and release guidance.

Generated-song playback can continue through the Studio transport, while Downloader and Stem Splitter results include direct audio previews before download.

This is a mobile web foundation, not a claim of a native app or offline audio processing. Rendering, downloading and remote separation still require network access.

Shared 48-Hour Result Download Window

IMPROVEMENTSTUDIO

Completed generated, humanized and mastered audio now uses one clear customer-facing rule: download it within at least 48 hours.

A scheduled cleanup removes stored audio on the first sweep after the deadline while preserving the account, payment and job record. Because cleanup runs on a schedule, removal can occur after—not before—the stated 48-hour window.

Unpaid source uploads use a separate seven-day window. Creators should still download and back up every completed file promptly rather than treating the Studio as permanent storage.

Free AI Music Export Inspector

FREE TOOLNEW FEATURE

The browser-based Export Inspector checks a finished song before a paid render, mastering session or distributor submission.

It reports container and codec information, sample rate, bit depth, sample peak, estimated true peak, integrated loudness, DC offset, mono compatibility, a frequency-ceiling heuristic and possible edge truncation.

Inspection stays in the browser tab. Measurements describe the file but cannot determine provenance, rights, platform acceptance or how a recording was made.