Story audio and voices · Chazz.ai

How story audio and voices works on Chazz.ai. Multi-voice narration, how speakers are detected, and exactly what it costs. A story can be narrated with a different voice per speaker. Text inside quotation marks is treated as speech and assigned to a character; everything outside them goes...

Multi-voice narration, how speakers are detected, and exactly what it costs.

A story can be narrated with a different voice per speaker. Text inside quotation marks is treated as speech and assigned to a character; everything outside them goes to the narrator.

How speakers are found

Attribution reads the speech verbs around each quote. The supported verbs are: said, asked, replied, whispered, shouted, exclaimed, muttered, growled, purred, stammered.

Where that is not enough, a separate model pass attributes the remaining lines. That pass is charged only when it actually runs.

What it costs

Narration runs on one of two voice planes, set per account in Settings → Voice quality. Standard is the default; Premium is ElevenLabs.

A character carries one voice on each plane, so switching quality changes the recording, not who they sound like.

The speaker-attribution pass adds a flat 10 CGC, charged only on runs where it happened. Source: worker/src/pricing.ts:106.

A single narration job is capped at 50,000 characters. Source: shared/constants.ts:921 (MAX_TTS_CHARACTERS).

Standard: 8 CGC
Premium (ElevenLabs): 80 CGC

Rough prices by length

Short: ~25 CGC standard / ~250 CGC premium
Medium: ~65 CGC standard / ~650 CGC premium
Long: ~120 CGC standard / ~1,210 CGC premium

Assigning voices

Regenerating audio is charged like a fresh generation. Get the voice assignment right before you press it.

A track is recorded on one plane end to end. Picking a voice from the other quality re-records the whole story rather than splicing two recordings together.

Save the story first: Speakers are detected from the saved prose.
Map each speaker: The voice dialog lists everyone it found. Assign a voice to each and one to the narrator.
Generate: The tracks are produced and stitched into one file with per-word timings, which is what drives the highlighting in the reader.

Elsewhere on Chazz.ai

This document is the server-rendered form of /docs/story-audio. The interactive page needs JavaScript, and shows the same content.

The complete documentation of this service is available as one plain-text file at /llms-full.txt. A short summary, with the operator's policy for automated readers, is at /llms.txt. Changes are published as a feed at /rss.xml.

Chazz.ai is operated by Cognitive Industries, ABN 62 794 528 747, Australia. Contact: [email protected].