Spoken translation: voice, pace and text-only cards
By default every card is read aloud in your output language, so your ears get the meaning while your eyes are on the words. You choose who reads it, how fast — and whether it is read at all.
The settings are in Settings then Session then Listening then Translation. Speaking has its own copy under Settings then Session then Speaking then Translation, so the two kinds can differ.
Choose a voice
Voice offers two:
- Marin — female (the default)
- Cedar — male
Press Preview next to the picker to hear a sample before you decide; press it again to stop. The voices are AI-generated.
Set the pace
Pace sets how fast the translation is read, from 0.5× to 1.5×. 1.0× is the voice’s natural pace. A change applies from the next sentence, so you can adjust it in the middle of a session. Double-click the slider to return to 1.0×.
- Slower (0.7–0.9×) helps when the output language is the one you are learning — for instance an English learner watching shows in their own language with English as the output.
- Faster (1.2–1.5×) keeps an evening of episodes moving once you mostly read the card.
Text-only cards
Turn Speak translations aloud off and cards arrive as text only: original, translation, vocabulary and grammar, with no spoken translation. The Translation replay button and the Voice and Pace settings disappear with it; the Original replay stays.
Two reasons to do it:
- It stretches your allowance. Generated speech is a large part of what a card costs, so text-only sessions use noticeably less of it. If you are near the end of your free credit or your month, this is the first switch to flip. How allowances work is on Plans and credit.
- It’s quieter. Late at night, in a library, or when you simply prefer to read.
Next
Skip ads, music and small talk — the suspend button.