Caption narration speed is automatically adjusted to fit within the time from when a caption starts until just before the next one begins (for the last caption in a step, it uses until the end of that caption). If the text is long relative to that time, the speed increases and narration becomes rushed. If it is hard to hear, review in this order:

Reference: Subtitles and text narration feature

Find captions that are being rushed

You can check which captions are being rushed in the Subtitle list of the video editor. Press Check narration speed, and any captions that do not fit the narration will display "rushed ×N speed", or a message indicating there are no rushed captions. After checking, if you change the timing of captions or pause durations of shapes, the display updates instantly. If you rewrite the text in a caption, check again.

The displayed multiplier reflects the speed after Solution 3 below (adding still duration in the step) returns speed to normal. Captions that have still duration and are actually read slowly are not counted as rushed.

Solution 1: Add space between captions

This is the most reliable approach. Delay the start of the next caption to create more time for narration.

Narration uses the time up until the next caption begins. So if you only extend the end of the current caption, without moving the start of the next one, the speed does not change. For the last caption in a step, narration can use until the end of the caption, so extend the end of that caption.

Solution 2: Break text into shorter sentences

If multiple sentences are in one caption, all of them must be read within that frame. Breaking text into separate captions, one per sentence, gives each its own frame and avoids crowding.

If you have already written the step text, you can select Create subtitles from the Subtitle list in the video editor, then choose From the step text, which creates captions where sentences are separated and assigned to step splits. For a single step, you can also create them from the button in the text field. Since the text is not squeezed into one frame, captions are less likely to rush.

Solution 3: Add shape still duration within the step

When you insert a shape (arrow, circle, spotlight, etc.), it adds a "Still Duration" where the video pauses. Narration does not pause during the still — it keeps reading, so the speed returns to normal for the duration of the still.

A caption that was rushed will be read slower by the amount of time gained from the still (but will not become slower than the original pace). Placing a shape where you want to emphasize caution also creates breathing room for narration.

Reference: Change "Still Duration" when inserting shapes (how many seconds to pause during playback)

Solution 4: Switch to text narration

If the step split is short and you cannot shorten the text or adjust the frame, you can use text narration instead of caption narration. Text narration does not sync with video, so no speed adjustment is applied and it reads at normal speed. The video repeats until narration finishes.

Change the narration target in Audio settings when you select the step in the video editor, switching between "Text" and "Subtitles" in Narration target. Since this is a per-step setting, you can use "Text" for just the steps that rush. The default when creating SOPs from materials is set in Step conversion settings.

Apply changes to existing SOPs

The behavior "narration does not pause during still duration" is already active in existing SOPs.

However, adjustments to the narration speed and volume themselves apply only when audio is recreated. Audio is recreated by operations like rewriting captions and saving, changing the original language, registering in pronunciation dictionary, or bulk-replacing text. Audio created earlier remains in its previous state until recreated.