Audio tags

Make hosts laugh, whisper, and react.

Audio tags are bracketed cues, such as a laugh, pause or whisper, that the Voice performs instead of reading aloud. In AI Podcasts Episodes you add them in the script editor; in Videos and Shorts, Jellypod writes them for you.

Which voices perform tags

Tags work only in speech blocks, with a Voice whose model performs them. A voice clone needs its Character’s Voice style set to More Expressive. For Voice Library voices, support depends on the Voice’s model, which the picker doesn’t show. Unsupported tags, and tags in a music block, are removed silently during audio generation.

The script editor shows a tag the Voice can’t perform as a warning-colored chip. Click it to open Audio Tags Unsupported, then choose Edit Character to pick another Voice or Clear Audio Tags to remove every tag in that block.

Insert a tag

1. Click into a speech block in the script editor.
2. Type [ or / to open the tag picker. It lists only the tags the block’s Voice performs. Keep typing to filter.
3. Click a tag, or use the arrow keys and press Enter or Tab. Escape closes the picker and keeps your text.

You can also type a full tag, such as [laughs]. Closing the bracket turns it into a chip:

“I just found out we hit a million downloads [laughs]. I honestly can’t believe it.”

Click a chip to select its text and type a replacement, or press Backspace or Delete over it to remove it.

Preset tags

Every Voice that performs tags offers the first column. Some Voices also perform the sounds in the second column, and the picker adds them for those Voices.

An emotion tag, dramatic or hesitates shapes every word after it until the end of the paragraph, for example [despair] I thought we had lost everything. [optimistic] Then your letter arrived. The other tags are a single sound at that point. Each Voice interprets tags differently, so listen again after changing Voices.

Custom tags

Type any short cue in brackets, such as [chuckles nervously] or [takes a deep breath]. The picker offers it under Custom. If the Voice misses it, rephrase it or use a preset.

What counts as a tag

Any bracketed text in a speech block is a tag, so a bracketed citation, timestamp or version number is stripped from the audio. Remove the brackets to have it read. Tags never appear in captions, SRT files or transcripts.

In an imported script, bracketed lines are read as speaker labels or folded in as cues; see Use your own script. To add recorded audio instead of a vocal effect, drag the file onto the timeline.

Backchannels

A backchannel is a short reaction, like “mhm” or “no way”, that the other Host says over the speaker. Write it between pipes inside the line it reacts to:

“We shipped the whole thing in a weekend |no way|, and nobody noticed until Monday.”

Type | partway through a line to open a picker of reactions, laughs such as |[laughs]|, and polite yields such as |sorry, go on|. Pick one, or keep typing your own and press Enter. The picker opens only when the script’s Hosts can perform Backchannels.

Backchannels play only with exactly two Hosts whose Voices both support them. Otherwise they are skipped silently, and the editor dims them (hover to see why). For two such Hosts, Jellypod writes its own Backchannels into generated scripts. Regenerating a block keeps its backchannels, and they never appear in captions or transcripts.

Audio tags in Videos and Shorts

The narration editor uses the same bracket syntax. Jellypod writes tags only when every selected Host’s Voice performs them. Moving a paragraph to a Voice that can’t turns its tags into warning chips.

Troubleshooting