RadioText AI：透過字幕提升網絡電台的無障礙體驗

RadioText AI: Making Internet Radio More Accessible

Hi! I'm Roman Eklmn, an independent developer building applications with AI under Eklmn Lab.

Today I'd like to introduce RadioText AI, my experimental Android application designed to make internet radio easier to follow through speech-to-text captions.

What does RadioText AI do?

RadioText AI lets you listen to internet radio while reading recognized speech on your Android screen.

The project is especially interested in accessibility for Deaf and hard-of-hearing users.

Current beta features include:

• Internet radio playback
• Speech-to-text captions
• Russian and English recognition
• Experimental detection of radio contest announcements
• Local caption history
• Android accessibility improvements

Why am I building it?

Radio is primarily an audio medium. I want to explore how speech recognition can help more people access spoken radio content.

I'm an independent developer, still learning, testing and improving the application.

Watch the demonstration

Download the Android beta

GitHub Releases:

https://github.com/OskolElkin/RadioText-AI-Releases/releases/tag/v0.2.0-beta

Alternative download on itch.io:

https://oskolelkin.itch.io/radiotext-ai

Current limitations

RadioText AI is a beta project, not a finished commercial product.

Recognition accuracy depends on audio quality, language, music, background noise and device performance.

Contest detection is experimental and may produce incorrect results. The app does not automatically enter contests.

Looking for volunteer testers

I'm looking for people willing to test RadioText AI on Android.

Feedback from Deaf and hard-of-hearing users would be particularly valuable.

I'd appreciate feedback about:

• Installation and first launch
• Caption accuracy and readability
• Ease of navigation
• Bugs and crashes
• Accessibility improvements

Please share your experience in the comments.

About the developer

Roman Eklmn — Eklmn Lab

GitHub: https://github.com/OskolElkin

Website: https://oskolelkin.github.io/eklmn-digital/

Thanks for checking out my project!