Skip to main content
Pass brand_glossary_id at the top level of the request, one glossary per call:
Video generation (studio, avatar and image videos, video batches per item, templates and Video Agent, set once per session) applies its pronunciations. Video translation, its batches per item, and proofreads apply its translation rules. Pronunciations apply where speech is synthesized from a script; your own audio_url or audio_asset_id plays as recorded. Video batch items are avatar or image. Glossary IDs are 32-character hex strings scoped to the workspace that owns them; an unknown ID fails with 400 before rendering, so no credits are spent. To shape how the video looks, pair the glossary with a brand kit on Video Agent.