Fablyn guide
Make your narration sound right
Fablyn’s narrators read your words exactly as you write them. Write for the ear, set the pace, and spell out the words people should hear.
Words are read as written
Numbers, dates and symbols are read the way you write them. Write “fifteen dollars” or “$15”, “7 am” or “seven in the morning”: both are read as written, so choose the form you want heard. Captions and the transcript show the same words.
Pauses come from punctuation
A comma is a short pause and a full stop a longer one. For an exact pause, end the line there and give the next one its own line: in Studio’s Narration tab, type the frames in Pause after under the line; an agent writes pauseAfterFrames.
Speed
Set the speed for the whole video, from 0.75× to 1.5×, and change it on any line that needs it. In Studio, open a Video, turn on Edit design and choose Narration. An agent writes "delivery": { "speed": 1.1 } on the document or on a line. A faster line is shorter, and the video follows it.
Pronunciation
Add a pronunciation to respell a word in what is spoken, never in the captions: “Fablyn” said as “Fab lin”, “SQL” as “sequel”. English entries match whole words. In Chinese, a pronunciation matches the characters wherever they appear, so write the shortest phrase that fixes the reading: 量两次, not 量, so that 数量 keeps its own reading. Each pronunciation applies once; a respelling is never respelled again. A pronunciation that is in no line changes nothing, and Studio says so.
What the narrator won’t read
The narrator treats anything in square brackets or braces as a stage direction and leaves those words out, and it reads “<|” and “|>” as its own tags. Fablyn refuses them in narration and offers commas instead: “Prices, in US dollars, start at fifteen.” Parentheses are fine.
Which narrator
Ethan, Hannah, Adrian and Freya narrate in English; Yutong, Zixuan and Xiaotong in Chinese. Each reads only its own language, so when you change a video’s language, choose one of its narrators. Play each one’s sample in Studio before you choose. Kore and the other Google voices have retired: videos they narrated still render from their recordings, and changing one asks you to choose a new narrator first.
What it costs
Each line is recorded once, then reused for free in every render until its words, speed or pronunciation change. New narration uses AI credits: 1.8 credits a minute of new audio. Hear this line in Studio records a line for real, so the next render reuses it at no cost. See the plan and AI credits.