@combycode/llm-sdk 1.6.0 → 1.7.0

package/dist/plugins/media/source-image.d.ts

@@ -26,6 +26,15 @@ export declare function toDataUrl(ref: NormalizedImageRef): string;
 export declare function openaiImageRef(ref: NormalizedImageRef): Record<string, string>;
 /** xAI image-ref object (`/v1/images/edits` image, video image). */
 export declare function xaiImageRef(ref: NormalizedImageRef): Record<string, string>;
+/** xAI video-ref object (`/v1/videos/extensions` + `/v1/videos/edits` `video`
+ *  field) — `{ url }` (public URL or base64 data-URL) or `{ file_id }`. Kept
+ *  separate from the image path so it doesn't run image mime-sniffing over
+ *  video bytes; the video mime is taken from the DataSource as declared. */
+export declare function xaiVideoRef(src: DataSource): {
+    url: string;
+} | {
+    file_id: string;
+};
 /** Google generateContent image part (inline base64 or Files-API file_uri). */
 export declare function googleImagePart(ref: NormalizedImageRef): Record<string, unknown>;
 /** Google Veo instance image (`:predictLongRunning` instances[].image).

package/dist/plugins/media/types.d.ts

@@ -18,6 +18,9 @@ export interface MediaMeta {
     durationMs?: number;
     sampleRate?: number;
     params?: Record<string, unknown>;
+    /** Provider-hosted URL (async video). Present when the bytes live remotely —
+     *  the browser renders from this since a cross-origin byte-fetch is CORS-blocked. */
+    sourceUrl?: string;
 }
 export interface MediaStore {
     save(id: string, data: Uint8Array, meta: MediaMeta): Promise<void>;
@@ -76,13 +79,22 @@ export interface VideoGenRequest {
     provider: string;
     model?: string;
     prompt: string;
+    /** First-frame image → image-to-video. */
     sourceImage?: DataSource;
+    /** Input video → extend or edit an existing clip (see `params.videoMode`).
+     *  The adapter routes to the provider's extension/edit endpoint instead of
+     *  plain generation. */
+    sourceVideo?: DataSource;
     params?: {
         duration?: number;
         aspectRatio?: string;
         resolution?: string;
         /** OpenAI Sora literal pixel `size` (e.g. "720x1280"). */
         size?: string;
+        /** When `sourceVideo` is set, which operation to run. `extend` (default)
+         *  continues the clip from its last frame; `edit` modifies it in place per
+         *  the prompt. Ignored without `sourceVideo`. */
+        videoMode?: 'extend' | 'edit';
     };
 }
 export interface MediaResult {
@@ -94,6 +106,12 @@ export interface MediaResult {
 export interface RawMediaResult {
     data: Uint8Array;
     mimeType: string;
+    /** Provider-hosted URL for the asset, when it exists (async video). In the
+     *  browser, cross-origin buckets (e.g. xAI vidgen) block a programmatic
+     *  byte-fetch via CORS, so `data` may be empty and this URL is the only way
+     *  to render (`<video src>` plays cross-origin without CORS) or to re-submit
+     *  as a `sourceVideo`. */
+    sourceUrl?: string;
     width?: number;
     height?: number;
     durationMs?: number;
@@ -116,6 +134,9 @@ export interface MediaCapabilities {
     audioGeneration: boolean;
     videoGeneration: boolean;
     audioStreaming: boolean;
+    /** Provider can extend/edit an existing video (`sourceVideo` on the request).
+     *  Undefined/false → passing `sourceVideo` throws. */
+    videoExtension?: boolean;
 }
 /** All MediaProviderAdapter HTTP calls now go through the NetworkEngine
  *  queue (rate limits, retries, hooks, observability) instead of holding a

package/dist/plugins/model-catalog/catalog.d.ts

@@ -39,6 +39,9 @@ export interface ModelCapabilities {
     imageGeneration: boolean;
     audioGeneration: boolean;
     videoGeneration: boolean;
+    /** Model accepts an existing video as input to extend/edit it (xAI
+     *  grok-imagine-video). Undefined/false → generation only. */
+    videoExtension?: boolean;
 }
 /** One generation parameter a media model accepts, with its allowed values.
  *  Either an enum (`values` + `default`) or a numeric range (`min`/`max`).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@combycode/llm-sdk",
-  "version": "1.6.0",
+  "version": "1.7.0",
   "description": "Unified, pluggable AI SDK for accessing the LLMs of every major provider (Anthropic, OpenAI, Google, xAI, OpenRouter) through one API. Cross-environment: Node, Bun, and the browser.",
   "type": "module",
   "main": "./dist/index.js",