Speech / Predict

• We've added a new endpoint which draws on 4 million data points to predict the duration of speech from a given script and list of voice names 🤓
• This will allow users to determine whether a voice is suitable, or whether their script is the correct length for their project, without needing to generate the audio itself, saving time and credits! ⏰
• Find out more and check out a code example in our guide, in the API reference or in the video below 👇

📢 We've added documentation and code examples on how to generate an audio "tag" and add it to the end of an existing audio advert. Find out more here.

Sound Templates

Sound template terms of use have been added for free and paying customers - we can now refer to the terms of use if people have questions about how they are allowed to use our templates.

You can refer here - https://audiostack.ai/terms-of-use

 Security updates

📘

Security updates

This doesn't mean that our old API was insecure, we're committed to updating and reducing any risks of compromise to our customers. We regularly review security updates, and we regularly patch any found errors.

We did a range of fixes to cover some security issues. These include

• https://nvd.nist.gov/vuln/detail/CVE-2022-25901
• https://nvd.nist.gov/vuln/detail/CVE-2022-25860
• https://nvd.nist.gov/vuln/detail/CVE-2021-44906
• https://nvd.nist.gov/vuln/detail/CVE-2023-26115

We are continuously working on these improvements and also decommissioned some older parts of our API as well.

Voice Playground (Platform) - Beta 🚧

• New workflow which is the replacement for the simple frontend we previously had in the example frontends
• A few small UI improvements and new sound library will come next week
• See it at http://platform.audiostack.ai

New Voice Library

• Updated the design
• Fixed provider names
• Is completely reusable and can be accessed at: http://library.audiostack.ai in the platform as a standalone workflow with the ability to preview, and in the voice select modal in the voice playground and sonic sell

SSML Prosody

We added the following functionality to have better prosody across providers. , ,

import audiostack


audiostack.api_key = "APIKEY"

text = """ The following part should be <as:prosody pitch="x-high" rate="x-fast" volume="x-loud">higher, faster and louder</as:prosody>,   
           """ 

script = audiostack.Content.Script.create(scriptText=text)
print(script.message, script.scriptId)
tts = audiostack.Speech.TTS.create(scriptItem=script, voice="henry", voiceIntelligence=True)
print(tts)
item = audiostack.Speech.TTS.get(tts.speechId)
item.download()           

Support for ElevenLabs v2 Multilingual Voices

You can now access v2 of ElevenLabs' multilingual voices via AudioStack, a major improvement in terms of improving content accessibility through language localisation. 🎉

Find out more about the capabilities of v2 on ElevenLabs' website.

We are changing the way we're providing sound templates to make it simpler for end users. Going forward, we're making it easier to understand what rights you have over the content you're working with, as we move towards offering a range of exclusive PRO-free sound templates (public rights organization free), which means that users can use the content accessed much more flexibly and with fewer restrictions.

More content is coming in the next few weeks, but while we move forward with this some of our old demo content, which had much more restrictive usage conditions, will be unavailable. If there's a sound template you really like and would like to use something similar for a broadcast ad or in a commercial setting, let us know and we will make sure we include a sound template that meets your needs.

Find out more about the sound templates we have available here:

<https://docs.audiostack.ai/docs/use-sound-templates>

Workflow Quick Start Guides

New quick start guides are now available in the AudioStack Docs, covering some of our users' most common workflows. You can find guides on:

• Creating an AI podcast and publish it using Julep
• Adding an AI voice-over to a video
• Using Dynamic Creative Optimisation to tailor audio ads to a specific customer

Are there other guides you'd like to see? Let us know in the discussion!

/TTS/Reduce

We've added a new endpoint to the AudioStack API, which allows users to condense their speech files to fit within a fixed duration, while retaining the speech's natural sound.

Common use cases include:

• 🕘 Creating audio adverts that need to fit within a fixed time slot (e.g. 15 or 30 seconds)
• 💬 Generating audio "tags" and terms and conditions text for the end of your audio file

Check out a code example and more info in our user guide.

Quick Start Guides

• We've made changes to our Quick Start guides, to help users get set up quicker. We now have two Quick Start guides - one aimed at people without software development experience, and a condensed version for developers. 📔
• These include updated code examples and step by step guidance on creating your first audio. 🔉

Mastering engine

We made a major update to our mastering engine, updating libraries and improving performance 🚀 You should have less error rates and a better customer experience.

Voice Intelligence Layer

• Voice Intelligence Layer Is the link to the voice intelligence layer check it out!
• Our celebrated voice intelligence layer is now UP TO 5X faster 🚤
• We've added to the voice intelligence layer an enhanced load balancer so now our platform works with longer form text. 💯
• This solves a major issue that many customers reported which was "fizzing out" of content. This should make your customers and users even happier 🐶

Error messages

• We've made many improvements to error messages so it should be clearer where some speech requests fail, and how to fix the error. This enhances the developer and user experience.