Capabilities
Call this endpoint before every creation flow:
GET /capabilitiesCapabilities describes products, file limits, video templates and voices, and image model options for the current account. Use it to build controls and validate requests instead of hard-coding a stale catalog.
Top-Level Sections
| Field | Meaning |
|---|---|
| products | The public products: video, image, and app |
| account | A safe account-plan summary |
| files | Supported file types, size, and count |
| video | Explainer Video options |
| image | AI Image model bundles |
| app | AI App options |
Explainer Video
video includes:
- duration_presets
- the currently fixed 16:9 aspect ratio
- languages and voices
- the current scene-template catalog
- independent expression styles
- supports_files, supports_url_context, and supports_messages
A template describes a use case and starting structure. An expression style describes how the video looks and communicates. Send template_id and expression_style_id independently.
AI Image
Each entry in image.models is an atomic model bundle. Current bundles are:
- gemini + nano-banana-2
- azure_openai + gpt-image-2
Each bundle advertises its own aspect_ratios, output_qualities, and qualities. Never mix a provider from one entry with a model or ratio from another.
output_quality is the delivered image size. quality is a GPT Image native setting and is valid only for azure_openai + gpt-image-2.
AI App
app lists supported languages and flags such as supports_files. AI App accepts uploaded files but does not currently advertise URL Context. Do not claim that it read a webpage merely because the prompt contained a URL.
Recommended Client Behavior
- Read capabilities when the creation screen opens.
- Cache briefly and support manual refresh.
- Refresh after account or plan changes.
- Display only options returned for the current account.
- Treat missing capabilities as unavailable instead of guessing.