Audiobook Studio 2.0 docs

Behavior Metadata

Instead of special-casing engines in Studio's code, engines declare their behavior as metadata. Studio reads it and adapts.

At a glance
  • Chunk limit and split target come from the manifest.
  • Progress is parsed from a declared pattern.
  • Settings render from a JSON Schema with UI hints.
  • Per-voice settings are declared, not hard-coded.

Text chunking

Each engine declares its text chunk limit. Studio splits long segments accordingly, so chunking adapts per engine without code changes.

Progress parsing

Progress comes from a declared pattern. Studio turns engine output into normalized progress events using metadata, not engine-specific parsing scattered through the app.

Schema-driven settings

Settings forms are generated from settings_schema.json, with optional UI hints. Engines can also declare per-voice settings. See manifest.json Reference.