Speak
Say the message the way you would normally say it.
Say what you mean in Amharic. QalbeQal gives you text you can review, copy, and use—without making you type the whole thought first.
Typing Fidel can turn one thought into a small project. QalbeQal gives you a shorter path without pretending the machine should speak for you.
Say the message the way you would normally say it.
Read the transcript. Fix a name or number if it needs you.
Copy it, insert it, or send it where the conversation already is.
QalbeQal turns your Amharic speech into Fidel. It can clean up spacing, spelling, and punctuation, but it should not summarize you, translate you, or replace your meaning with its own.
Use the iPhone app, the Mac shortcut, or the Telegram bot. The path changes with the device; the job stays the same: get your Amharic out of your head and into text.
Add the people, places, and work words you use often. QalbeQal uses those terms as context so it has a better chance of writing them correctly—including the English that naturally shows up in Amharic speech.
You should not need a studio. A few simple habits give any speech recognizer a cleaner chance to hear you.
Use your normal voice. Clear sentences work better than whispering, shouting, or speaking one word at a time.
Move away from TV, music, and wind. You do not need silence—just make your voice the clearest sound.
Overlapping voices are hard to separate. Let one person finish before another starts when the transcript matters.
Check names, numbers, dates, and addresses. Add repeated special terms to your personal dictionary.
From family messages to work, school, faith, and community—speak the thought and turn it into Fidel you can use.
Say the full thought instead of fighting a keyboard halfway through it.
Draft notes, updates, and ideas while the words are still fresh.
Capture announcements, reflections, and the language people actually use together.
Give people another way to write when Fidel input is slow, unfamiliar, or physically difficult.
QalbeQal processes audio to give you a transcript. It does not sell your data, follow you around the internet, or quietly turn ordinary dictation into training material.