This week we worked hard on improving our SonicSell experience based on customer feedback. All these little changes every week help us move the product forward. :100:

We also shipped a major improvement to our copy generation and we shipped stem separation.

Stem Separation

Improvements and a lot of bug fixes

  • We shipped some major improvements to copy generation in SonicSell - so your advert should now have significantly less deviations. We also added retry logic and better examples through improved prompt engineering. :100:
  • :+1:Update success/error toasts file uploads
  • :thumbsup:SonicSell Header improvements
  • :+1:Remove credit-counter (users complained it was distracting)
  • Powered by AudioStack is now clearer
  • :bug: Fix Workflow Inconsistencies in SideBar
  • :bug: Support special chars in files table
  • :bug: Updated copywriting to enhance the user experience

Coming soon :construction:

  • We shipped some great stuff in the backend too, and we'll speak more about that next week :100:
  • We also invested in our ad-serving capabilities and analytics capabilities - please reach out if you want to know more about this.

This week, we've added a few UX improvements to make your experience of AudioStack Platform and workflows smoother.

New ad player in SonicSell

  • You can now give feedback (thumbs up or down) about the quality of ads in SonicSell, to help improve the content you hear in future.

Platform improvements

  • Added a new parrot to the favicon :bird:
  • New workflow cover images on Home page
  • New Navigation - easier to find your workflow

Files Page improvements

  • File view made into page - this makes it easier to use on mobile and allows us to add exciting new features. :eyes: Watch this space!

Voice Cloning improvements

  • You can now add your own alias for new voices instant cloning voice to make it easier to find in your workflows.
  • Integrated new voice builder in the platform to deprecate the previous one.

Denoise your noisy audio files

Docs improvements

We added some improvements to improve the onboarding experience :100:


  • We added zip downloader added to the api for exporting zips :zap:


  • Users can join a Waitlist for the ‘Coming Soon’ Workflows
  • Implement security headers into nextjs apps enhancing our security :lock:
  • Users can now brand the Platform with their logo - a much requested feature
  • Audio title encoding in the links to player doesn’t support special chars :bug:
  • The voice library now shows the best voices at the top


Check out the library

You can check it out at our Library page!


  • AdCards now provide better error feedback for invalid input


  • Bug fixes in the voice cloning experience :bug:

Platform improvements

  • Sound Library now shows the sound segments that belong to each template
  • Sound Library page layout is now improved to have more consistent sizing
  • Added Denoise and Source separation functionalities to the File Edit modal
  • Clicking on the sidebar header now redirects to home
  • Updated link to contact about enterprise plans


  • SonicSell now catches bad product descriptions and errors appropriately.
  • SonicSell now calls /content/recommend/ to get better tone and mood selections via ms--text-tagger (ML endpoint fully integrated into a workflow and charging credits)
  • We enhanced our LLM integrations and improved the workflow

New Features

  • Source separation and better performance - this means that you can now upload an audio file containing both speech and music, and separate the voice out from the music to use in your workflows (such as voice cloning or sound template creation).

New voice builder in production!

What's new?

  • central service for all things voice cloning - easy integrations of new vc providers
  • possibility of cloning using multiple files
  • fixes a bunch of issues we've had so far e.g. timeouts, file errors
  • Adding more features , (SSML, STS in API, fast synthesis)
  • internal file processing to make sure user's input matches provider's requirements
  • internal tool - python package useful for experiments and testing integrations


  • Shipped some small improvements to file managements, i.e. better validation on broken files, and better performance :100:
  • shipped the capacity of using any file id as sound template in mastering :tada:
  • Improvements in the speed and reliability of AES :100:
  • mastering improvements for bugs with not file errors :minidisc:

 Login improvements

:rocket: By clicking on the log in button, you will now be redirected to log in to the AudioStack platform, making our documentation easier to navigate. :+1:

Python SDK enhancements.

  • Improve messages in authentication errors
  • Added tags to the SDK
  • And some improvements to error messages and quality of life improvements

Have a look here Python SDK

Organisations containing non USASCII characters in their names

:bug: Organisations which contain a non USASCII char (e.g. ä,ö,ñ) couldn't login to the platform at all. :boom: The encoding broke and the Organisation name was invalid.

Now this is fixed :rocket:

Fantastic work by our team. Please us ALL THE £$%^`6())( that you want :100:

Signing in using SSO

  • We fixed a bug that was blocking people from signing up to SSO.

Cached files

:bug: We fixed a bug where the User tried to request a cached file with a public accessible URL and instead get a private URL back. Cached files always returned private urls even the User explicit want a public one.

🎉 Exciting news 🎉

Did you know you can now create your own sound template from the AudioStack platform? Simply navigate to the Content/ Resources page, upload a file and follow the instructions in the docs to create a private sound template:

You will be able to use your sound design in any of our workflows, including SonicSell.

We'll soon be updating the production credits charged for interacting with endpoints of AudioStack API, along with AudioStack platform functionality that uses the API. These changes will result in increased pricing in some areas and decreased pricing in others, to better reflect the complexity of the functionality.

The production credits being updated are:

If you have any queries, please contact our team at [email protected].

We have added audio enhancement endpoints, added our Text Tagger to the SDK and added some security improvements and API improvements for our Voice Intelligence Layer

New Audio enhancement endpoints

We are continuously working on enriching our audio capabilities so we've added the ability to evaluate, enhance and denoise audio files.

:headphones: We added evaluate audioEvaluate an audio file.

:headphones: Enhance audio Evaluate and enhance an audio file.

:headphones: Denoise an audio file in order to remove background noise. Denoise an audio file.


:100: A quality of life improvement the Voice Intelligence Layer is now enabled if you use one of our in built tags such as <as:break>

:100: We also added improvements to the file transfer routes so you no longer need to add file names

:bug:We added some security improvements for our customers to the voice preview

Recommendations using our Text Tagger

We've added recommendations for tags, moods and tones using the Text Tagger

See how to use them on our docs -

text = "AudioStack's technology seamlessly integrates into your product or workflow and cuts your audio production cycles to seconds while making your budgets go further."
number_of_results = 2
tags = ["happy", "fun", "valuable", "fast"]
category = "my category"

tag_item = audiostack.Content.RecommendTag.create(
    text=text, number_of_results=number_of_results, tags=tags, category=category

mood_item = audiostack.Content.RecommendMood.create(
    text=text, number_of_results=number_of_results

tone_item = audiostack.Content.RecommendTone.create(
    text=text, number_of_results=number_of_results

Another exciting addition to the AudioStack API - the ability to evaluate a given audio file to receive a quality evaluation, and then to enhance or denoise your audio file as needed.

We added three new endpoints as part of the Production Suite (i.e. /production/suite/), which can help you to check the quality of your recording, for example before using it for voice cloning.

With the /production/suite/enhance endpoint, you can let AudioStack do the work and both evaluate and enhance your file as needed.

Find out more in the docs: