Listen Anywhere: Your Articles Read Aloud by AI

Life is a blur of constant activity, and we often find ourselves caught up in the hustle and bustle. From crowded subways to busy kitchens and long waits at traffic lights, reading has become a small luxury squeezed into the spare moments of our day.

Our eyes are tired, yet the list of unread articles keeps growing. Maybe it’s time to close your eyes and let the words flow right to your ears.

Cubox now supports AI text-to-speech on iPhone and iPad apps. Simply open an article you've saved, tap the play button in the bottom menu, and before you know it, you’ve listened to the entire thing.

Cubox is different from a typical book or podcast app. With so many new articles being saved every moment, it's impossible for us to pre-process a massive volume of audio conversions. To address this, we've created a custom solution tailored for the "read-later" experience, offering a range of features that work right out of the box:

• Multiple natural AI voices are available and can be switched in real time.
• The AI voices support both Mandarin and English.
• You can choose from eight different playback speeds and change them on the fly.
• The current sentence is highlighted with an underline as the audio plays.
• Tapping the player's title takes you back to that spot in the article.
• An estimation algorithm allows you to seek forward and backward in the audio.
• You can play and control the audio globally within the app while performing other actions like highlighting.
• The player works in the system background, so you can listen while using other apps.

All these features are designed with the same intuitive feel as your favorite system media player. We're confident you'll know how to use them right away, so we won't go into more detail. Feel free to update your app on the App Store and try it out.

In a previous update, we introduced a new option in the unread status settings. You can now automatically mark an article as "read" once you've finished reading or listening to the whole thing. This way, you don't have to manually mark it as read after listening.

Want to know more details? Read on for our FAQ section below.

Frequently Asked Questions

What languages are supported?

Currently, we support Mandarin and English. Note that the voice options will change based on your Cubox app language setting. For example, if you primarily read and listen to content in Chinese, you can go into Cubox's advanced settings and set the app language to Chinese to see a variety of Mandarin voice options, regardless of your device's default language or region.

If you need to listen to both Chinese and English content, the last two voice options in the Chinese list, which have English names, are optimized for reading pure English text.

What kind of content can be read?

The feature supports articles and snippets, including newsletters. This is content collected from external sources with good parsing quality, which makes it suitable for voice synthesis.

Is there a feature to generate AI podcasts?

Not yet. Instead of focusing on podcasts, we're exploring how a large AI model could provide a more efficient "audio summary" that distills the core ideas of an article. We believe that even the best current AI podcast solutions can cause listening fatigue over time. Spoken language is a unique art form, and there are already plenty of great podcasts out there—it's not necessary for every article to become one.

Will this be available on Android?

The Android version is currently in development and will be released soon.

Can I start reading from a specific point in the article?

Currently, you cannot select a specific point in the text to begin playback. However, articles are usually not very long, and you can use the fast-forward and rewind functions with the text as a guide to adjust the playback position.

Why isn't the progress bar always accurate?

Since the AI voice is generated in real time, the playback progress and time are estimates and may differ from the actual duration. They are meant to be used as a reference.

What AI voice synthesis model do you use?

Cubox currently uses Unreal Speech for English. The voice list will match your current Cubox app language setting. We're considering adding intelligent language switching and support for more large language models in the future.

How much can I listen to?

Based on our system, 500 characters of playback consumes one credit. The newly upgraded monthly plan of 2,000 credits allows you to listen to about 100 long-form articles (over 10,000 characters each) per month. As with our previous AI upgrades, we'll continue to balance usage and cost and may explore other solutions to increase the available listening time.

Why did it take you so long to add this feature?

Before modern AI models were available, traditional text-to-speech had a very robotic quality. While we’ve been getting requests for this feature for years, we waited until now because we finally feel the quality is good enough to integrate.

Why should I listen to articles in Cubox?

There are many tools for converting text to speech or podcasts. But Cubox's design isn't about a single feature; it's about providing a space that connects your entire reading workflow. From saving and archiving to reading and highlighting, Cubox makes it easy for serious readers to manage everything more seamlessly than with other note-taking or knowledge management apps.

If you have other suggestions or questions, the best way to reach us is by emailing hi@cubox.cc. Thank you!