An older television series was being prepared for rerelease, but its original music could not simply be removed. Dialogue, laughter, sound effects, and the score had all been mixed into one audio track, and the original stems were no longer available.
Jaywalker Music used AudioShake to separate the old music while preserving the other sounds, then placed a new score into the program. According to a customer case published by AudioShake in December 2025, the workflow helped replace more than 200 tracks whose licenses had expired, allowing the programs to become eligible for distribution again. This is a vendor-published customer account, not an independently measured result. Jaywalker use case
While much of the AI music market focuses on generating new songs, AudioShake works in the opposite direction: it takes sound that has already been mixed and separates it into editable parts.
The company says this has become more than a technical demonstration. In October 2025, AudioShake reported nearly 400% year-over-year revenue growth, more than 100 million minutes of audio processed during the previous year, and over 40 enterprise partners. The company did not disclose its absolute revenue, and the figures were company-reported rather than independently audited. Company growth disclosure
The buyer is not paying for another song. It is paying to keep content it already owns usable, editable, and distributable.
Remove the music, but keep the laughter
AudioShake is an AI sound-separation system. It can split a song into vocals and instruments, and it can separate film and television audio into dialogue, music, and effects that post-production teams can edit independently. Film and television product description
For Jaywalker, the problem was not simply muting the audio. Turning off the entire track would remove the background music, but it would also erase dialogue, laughter, and reactions recorded on set. The customer needed to replace only the music while preserving as much of everything else as possible.
AudioShake’s product page provides three versions of the same broadcast sample: the original mix, a version with the music removed, and a version with replacement music. The clips below use the same first 12 seconds so the difference can be heard directly.
Original mix
Music removed
Replacement music
Source: three versions of the same official sample on AudioShake’s film and television product page.
Removing a music track does not automatically resolve copyright questions. Jaywalker still had to create replacement music, and the rights holders still had to license and distribute the new version. AI handles the difficult intermediate step of separating material that had previously been locked together.
Change the language without rebuilding every sound
CrunchLabs, the science-video company founded by Mark Rober, faced a related problem.
The team wanted to create multilingual versions of its existing videos, but some older content was available only as finished mixes. Adding a new voiceover on top of the original dialogue would cause the voices to compete. Removing all of the original sound would also remove experiment noises and the participants’ reactions.
According to a customer case published in December 2025, CrunchLabs uses AudioShake to separate dialogue, music, and sound effects before producing different language, sponsor, and regional versions. It can also remove commercially licensed music that is unsuitable for a target market and replace it with another track. CrunchLabs customer case
One detail explains why separation matters. Nonverbal reactions such as laughter and surprise can remain in the background and effects tracks. The spoken language can change without replacing the authentic reaction when an experiment succeeds.
This is different from merely generating an AI voiceover. New dialogue is only one part of the delivery. Deciding which elements of the original production should change and which should remain also determines whether the finished version is usable.
From editing content to operating inside another product
In September 2026, AudioShake moved into a workflow that sits farther from creative production and closer to rights reporting: documenting music usage.
Music Reports integrated AudioShake into its Cuetrak platform. A production company can upload a media file, after which the system separates the music, identifies tracks and their timing and duration, and passes that information to Cuetrak as a cue sheet for human review. Music Reports integration announcement
A cue sheet is not a consumer feature. It supports rights reporting and later royalty settlement. Music that is covered by dialogue, commentary, or crowd noise can be difficult to identify. AudioShake’s role is to isolate that music before the identification and reporting steps.
The company does not have to persuade customers to replace their existing software. It can become one step inside software they already use.
Existing Cuetrak customers can contact Music Reports to enable the upload workflow without first adopting AudioShake as a standalone end-user product.
That is a different sales position from offering a single stem-separation button. A standalone tool depends on users remembering to open it. An integration appears inside a task the customer already has to complete.
The announcement does not disclose contract value, adoption, or attributed revenue. It shows that the capability is available in the workflow, not that every Cuetrak customer has activated it or that cue-sheet production is fully automatic.
Annual access plus usage-based charges
AudioShake described its enterprise model in October 2025 as an annual access fee combined with usage-based pricing. At the time, it named customers including Universal Music Group, Disney Music Group, Warner Music Group, and NFL Films. The disclosure does not provide price levels or establish that every named organization buys the same package. Enterprise pricing model and customer disclosure
The model matches the work. An enterprise buyer is not separating one song as an experiment. It may have a television library, a catalog of music, or a continuing stream of new media that must be processed. Renewal depends on whether the output can enter editing, localization, reporting, and delivery workflows, not only on whether a first demonstration sounds impressive.
AudioShake released a real-time processing SDK in June 2025 across iOS, macOS, Android, Windows, and Linux. In December of that year, it expanded music removal into live and streaming use cases. The 2026 Cuetrak integration then connected separated audio with identification and reporting. Real-time SDK announcement
The progression offers a practical product lesson. After audio is separated, an editor needs to replace music, a distribution team needs localized versions, and a rights team needs a usage report. Each additional downstream job gives sound separation another opportunity to become part of recurring production rather than a tool that users try once and leave.