@tanstack/ai 0.52.0 → 0.52.2

package/skills/ai-core/media-generation/SKILL.md

@@ -521,23 +521,26 @@ const { jobId } = await generateVideo({
 // (x-goog-api-key header or ?key= query parameter).
 ```
 
-Gemini Omni Flash (`geminiVideo('gemini-omni-flash-preview')`) is served by
+Gemini Omni Flash (`geminiVideo('gemini-omni-1.1-flash')`) is served by
 the Interactions API instead of Veo's operations flow — same adapter, routed
-by model. Clips are 720p; `duration` is any number of seconds in the 3–10
+by model. `duration` is any number of seconds in the 3–10
 range (fractional ok, default 10 — availableDurations() reports the range),
-`size` is the aspect ratio (`'16:9' | '9:16'`), and the finished video arrives
+`size` is an `aspectRatio_resolution` template (`'16:9'` or `'16:9_1080p'`;
+suffix `'360p' | '720p' | '1080p' | '4k'`, default 720p), and the finished video arrives
 **inline** as a `data:video/mp4;base64,…` URL (no key needed to use it).
 Image/video prompt parts are sent as interaction content blocks, grouped
 as images, then videos, then text (no
 `metadata.role` routing); `data` sources go inline, `url` sources pass
 through as-is (never downloaded — use Gemini Files API URIs for remote
 media). For conversational editing, pass a prior generation's `jobId` as
-`modelOptions.previous_interaction_id` with a prompt describing the change:
+`modelOptions.previous_interaction_id` with a prompt describing the change.
+`gemini-omni-flash-preview` remains a deprecated alias until it shuts down
+on 2026-09-30.
 
 ```typescript
 import { geminiVideo } from '@tanstack/ai-gemini'
 
-const omni = geminiVideo('gemini-omni-flash-preview')
+const omni = geminiVideo('gemini-omni-1.1-flash')
 const first = await generateVideo({
   adapter: omni,
   prompt: 'A violinist outdoors',

package/src/activities/chat/index.ts

@@ -623,12 +623,12 @@ export function createChatOptions<
     TStream,
     InferredContext<TTools, TMiddleware>
   >,
-  'tools' | 'middleware' | 'interrupts'
+  'tools' | 'middleware' | 'interrupts' | 'context'
 > & {
   tools?: TTools
   interrupts?: TInterrupts
   middleware?: ExactMiddlewareOption<TTools, TContext, TInterrupts, TMiddleware>
-} {
+} & RuntimeContextOption<TTools, TMiddleware, TContext> {
   return options
 }
 

package/src/activities/chat/stream/processor.ts

@@ -1357,6 +1357,12 @@ export class StreamProcessor {
       state.hasToolCallsSinceTextStart = false
     }
 
+    // A segment begins at its first delta, not its second. Left set, the flag
+    // makes the second delta look like a further segment: the accumulation
+    // resets and updateTextPart writes over the part the first delta created.
+    // A later tool call sets it again in handleToolCallStartEvent.
+    state.hasToolCallsSinceTextStart = false
+
     const currentText = state.currentSegmentText
     const delta = chunk.delta || ''
     const nextText = delta !== '' ? currentText + delta : currentText