Hi HN,
I'm P.F. Tuan, an indie, and the solo developer behind AI Voice AudioBook. Thanks for checking out my project!
The "Why":
I have a long commute and a huge list of articles and books I want to get through. I tried a bunch of text-to-speech apps, but I could never stick with them because the robotic, generic voices were so grating to listen to for more than a few minutes.
I started wondering: what if I could listen to a book narrated in a voice I actually enjoyed, or even my own? That's the core idea behind this app.
The "What":
AI Voice AudioBook lets you turn any ebook (EPUB, PDF, TXT) or web article into an audiobook. The key feature is the voice:
You can clone your own voice by providing a short audio sample.
Or, you can use one of the premium, natural-sounding AI voices available in the app.
It also automatically strips out ads and other clutter from web pages, so you get a clean listening experience.
Tech Details:
The app is built natively with Flutter. For the backend, the voice cloning and TTS generation are handled by a combination of commercial and fine-tuned open weight models based on self-collective dataset. One of the trickiest parts was reliably parsing clean text content from messy HTML, which leverages LLM processes.
What I'm looking for:
I'd be incredibly grateful for your feedback, especially on a few things:
The Voice Cloning Process: Is it intuitive? How is the quality of your cloned voice for long-form listening? Especially for non-English languages.
Is the service stable for various texts? Especially for non-English languages, we need to adjust our training dataset according to the feedback.
Is the UI convenient? As there are many AI tools inside.
The Web Article Parser: Does it work well on your favorite sites?
Pricing: The AI models are expensive to run. Does the current pricing structure seem fair and make sense to you?
I'll be here all day to answer every question. Thanks again for your time!
P.F. Tuan from ZanChat
