chore (ai): stable sendStart/sendFinish options (#6499)

Author: lgrammel

.changeset/rare-foxes-build.md

@@ -0,0 +1,5 @@
+---
+'ai': major
+---
+
+chore (ai): stable sendStart/sendFinish options

examples/next-openai/app/api/use-chat-custom-sources/route.ts

@@ -27,11 +27,7 @@ export async function POST(req: Request) {
         messages: convertToModelMessages(messages),
       });
 
-      writer.merge(
-        result.toUIMessageStream({
-          experimental_sendStart: false,
-        }),
-      );
+      writer.merge(result.toUIMessageStream({ sendStart: false }));
     },
     originalMessages: messages,
     onFinish: options => {

packages/ai/core/generate-text/stream-text-result.ts

@@ -82,7 +82,7 @@ export type UIMessageStreamOptions = {
    * that send additional data.
    * Default to true.
    */
-  experimental_sendFinish?: boolean;
+  sendFinish?: boolean;
 
   /**
    * Send the message start event to the client.
@@ -95,7 +95,7 @@ export type UIMessageStreamOptions = {
    * streamText calls that send additional data to prevent
    * the message start event from being sent multiple times.
    */
-  experimental_sendStart?: boolean;
+  sendStart?: boolean;
 
   /**
    * Process an error, e.g. to log it. Default to `() => 'An error occurred.'`.

packages/ai/core/generate-text/stream-text.test.ts

@@ -999,9 +999,7 @@ describe('streamText', () => {
         ...defaultSettings(),
       });
 
-      result.pipeUIMessageStreamToResponse(mockResponse, {
-        experimental_sendFinish: false,
-      });
+      result.pipeUIMessageStreamToResponse(mockResponse, { sendFinish: false });
 
       await mockResponse.waitForEnd();
 
@@ -1447,9 +1445,7 @@ describe('streamText', () => {
         ...defaultSettings(),
       });
 
-      const uiMessageStream = result.toUIMessageStream({
-        experimental_sendFinish: false,
-      });
+      const uiMessageStream = result.toUIMessageStream({ sendFinish: false });
 
       expect(
         await convertReadableStreamToArray(uiMessageStream),
@@ -1472,9 +1468,7 @@ describe('streamText', () => {
         ...defaultSettings(),
       });
 
-      const uiMessageStream = result.toUIMessageStream({
-        experimental_sendStart: false,
-      });
+      const uiMessageStream = result.toUIMessageStream({ sendStart: false });
 
       expect(
         await convertReadableStreamToArray(uiMessageStream),

packages/ai/core/generate-text/stream-text.ts

@@ -1411,8 +1411,8 @@ However, the LLM results are expected to be small enough to not cause issues.
     messageMetadata,
     sendReasoning = false,
     sendSources = false,
-    experimental_sendStart = true,
-    experimental_sendFinish = true,
+    sendStart = true,
+    sendFinish = true,
     onError = () => 'An error occurred.', // mask error messages for safety by default
   }: UIMessageStreamOptions = {}): ReadableStream<UIMessageStreamPart> {
     const lastMessage = originalMessages[originalMessages.length - 1];
@@ -1539,7 +1539,7 @@ However, the LLM results are expected to be small enough to not cause issues.
             }
 
             case 'start': {
-              if (experimental_sendStart) {
+              if (sendStart) {
                 const metadata = messageMetadata?.({ part });
                 controller.enqueue({
                   type: 'start',
@@ -1551,7 +1551,7 @@ However, the LLM results are expected to be small enough to not cause issues.
             }
 
             case 'finish': {
-              if (experimental_sendFinish) {
+              if (sendFinish) {
                 const metadata = messageMetadata?.({ part });
                 controller.enqueue({
                   type: 'finish',
@@ -1587,8 +1587,8 @@ However, the LLM results are expected to be small enough to not cause issues.
       messageMetadata,
       sendReasoning,
       sendSources,
-      experimental_sendFinish,
-      experimental_sendStart,
+      sendFinish,
+      sendStart,
       onError,
       ...init
     }: ResponseInit & UIMessageStreamOptions = {},
@@ -1602,8 +1602,8 @@ However, the LLM results are expected to be small enough to not cause issues.
         messageMetadata,
         sendReasoning,
         sendSources,
-        experimental_sendFinish,
-        experimental_sendStart,
+        sendFinish,
+        sendStart,
         onError,
       }),
       ...init,
@@ -1625,8 +1625,8 @@ However, the LLM results are expected to be small enough to not cause issues.
     messageMetadata,
     sendReasoning,
     sendSources,
-    experimental_sendFinish,
-    experimental_sendStart,
+    sendFinish,
+    sendStart,
     onError,
     ...init
   }: ResponseInit & UIMessageStreamOptions = {}): Response {
@@ -1638,8 +1638,8 @@ However, the LLM results are expected to be small enough to not cause issues.
         messageMetadata,
         sendReasoning,
         sendSources,
-        experimental_sendFinish,
-        experimental_sendStart,
+        sendFinish,
+        sendStart,
         onError,
       }),
       ...init,