- Urdu and English in one sentence: how well does dictation cope?
'Kal meeting hai at 3 pm, please remind me.' This is how we talk.
8 - What would make you recommend XIRO to a friend?
Mine would be: it just works on the first day. What is your bar?
32 - Indonesian and Javanese: regional speech support
Many families speak both. Plans?
18 - Accents: my assistant keeps mishearing my Nigerian English
It is fine with news-reader English and shaky with real life.
9 - Text-to-speech in Urdu: which voices feel natural?
Naturalness matters more than clarity for long reading.
11 - Brazilian Portuguese vs European Portuguese voices
The voices and spellings differ a lot. Which are you using?
5 - Arabic dialects: Egyptian, Gulf, Levantine and the Fusha problem
Formal Arabic works. Real speech is a different story.
5 - Whisper on noisy audio: pre-processing that actually helps
Noise suppression helped some, hurt others.
5 - Is voice AI still fundable after the hype?
Investors want revenue now, not demos.
18 - Hindi and Hinglish: what works, what breaks
Script mixing is the main headache.
5 - Japanese politeness levels in assistant replies
Keigo, polite, casual? A setting for formality would be ideal.
6 - Wake words in non-English languages
Do they work as well as English ones?
6 - Right-to-left display in the HUD: fixes and gaps
Numbers inside Arabic text still jump around.
4 - Learning Spanish with a voice partner: a routine that works
Daily ten minutes of corrected conversation.
5 - Sign language and voice: can assistants bridge both?
Camera-based sign recognition plus speech could connect communities.
5 - Which language should XIRO add next? 🗳
Vote with a reply. I'll count the votes at the end of the week.
6 - Developer API wishlist: what should XIRO expose?
Events, actions, memory read access, and a plugin manifest?
6 - German compound words and dictation errors
'Geschwindigkeitsbegrenzung' tests every system 😅
7 - French liaison and recognition: why 'les amis' trips models
Silent letters and liaison confuse transcription.
11 - Children's voices and elderly voices: recognition gaps
Models trained on adult voices struggle at both ends.
4