LiveVideoGenerationResult

Defined in: packages/ai/src/types.ts:2327

Experimental

Result of live generation. JSON-serializable so a server route can return it to a browser.

Reactor: connect with token and model (the connect slug). fal: model is the WMA app id. Open wma(model) through a server proxy that attaches FAL_KEY. Do not send token as Key credentials.

Live generation is an experimental feature and may change.

Properties#

expiresAt#

expiresAt: number;

Defined in: packages/ai/src/types.ts:2338

Experimental

Token expiry as milliseconds since epoch

id#

id: string;

Defined in: packages/ai/src/types.ts:2329

Experimental

Unique identifier for this generation

model#

model: string;

Defined in: packages/ai/src/types.ts:2334

Experimental

Connect id for the browser client. Reactor: reactor/helios. fal: WMA app id fal-ai/minimax-h3-max-director.

prompt#

prompt: string;

Defined in: packages/ai/src/types.ts:2340

Experimental

Prompt the client should send when it starts the session

sessionId?#

optional sessionId?: string;

Defined in: packages/ai/src/types.ts:2344

Experimental

Provider session id, when the adapter created one

status#

status: "ready" | "waiting";

Defined in: packages/ai/src/types.ts:2342

Experimental

Session status after the server half finishes

token#

token: string;

Defined in: packages/ai/src/types.ts:2336

Experimental

Short-lived session token. Reactor uses this to connect. fal does not.

usage?#

optional usage?: TokenUsage<ProviderUsageDetails>;

Defined in: packages/ai/src/types.ts:2346

Experimental

Token usage / billing, when the adapter can report it