fix(provider/groq): experimental_transcribe fails with valid Buffer (#8159)

## Background

`experimental_transcribe` fails with valid Buffer input after upgrading
to [email protected]

## Summary

[Set
`filename`](https://developer.mozilla.org/en-US/docs/Web/API/FormData/append#filename)
when uploading an audio blob to Groq's transcribe API

## Manual Verification

- https://github.com/firstloophq/ai-sdk-audio-bug-repro

## Tasks

<!--
This task list is intended to help you keep track of what you need to
do.
Feel free to add tasks and remove unnecessary tasks or this section as
needed.

Please check if the PR fulfills the following requirements:
-->

- [x] Tests have been added / updated (for bug fixes / features)
- [x] Documentation has been added / updated (for bug fixes / features)
- [x] A _patch_ changeset for relevant packages has been added (for bug
fixes / features - run `pnpm changeset` in the project root)
- [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the
project root)

## Future Work

Apply the same fix to other places where we upload a blob without
defining a file name

```
packages/elevenlabs/src/elevenlabs-transcription-model.ts:
  75      formData.append('model_id', this.modelId);
  76:     formData.append('file', new File([blob], 'audio', { type: mediaType }));
  77      formData.append('diarize', 'true');

packages/gladia/src/gladia-transcription-model.ts:
  502        'audio',
  503:       new File([blob], 'audio', { type: options.mediaType }),
  504      );

packages/openai/src/transcription/openai-transcription-model.ts:
  131      formData.append('model', this.modelId);
  132:     formData.append('file', new File([blob], 'audio', { type: mediaType }));
  133  

packages/revai/src/revai-transcription-model.ts:
  254  
  255:     formData.append('media', new File([blob], 'audio', { type: mediaType }));
  256      const transcriptionModelOptions: RevaiTranscriptionAPITypes = {
```

## Related Issues

Closes #6413
Closes #7757

Author: gr2m

.changeset/smooth-impalas-remain.md

@@ -0,0 +1,6 @@
+---
+'@ai-sdk/provider-utils': patch
+'@ai-sdk/groq': patch
+---
+
+fix(provider/groq): `experimental_transcribe` fails with valid Buffer

examples/ai-core/src/transcribe/groq-string.ts

@@ -7,6 +7,11 @@ async function main() {
   const result = await transcribe({
     model: groq.transcription('whisper-large-v3-turbo'),
     audio: Buffer.from(await readFile('./data/galileo.mp3')).toString('base64'),
+    providerOptions: {
+      groq: {
+        responseFormat: 'verbose_json',
+      },
+    },
   });
 
   console.log('Text:', result.text);

packages/groq/src/groq-transcription-model.ts

@@ -6,6 +6,7 @@ import {
   combineHeaders,
   convertBase64ToUint8Array,
   createJsonResponseHandler,
+  mediaTypeToExtension,
   parseProviderOptions,
   postFormDataToApi,
 } from '@ai-sdk/provider-utils';
@@ -68,7 +69,12 @@ export class GroqTranscriptionModel implements TranscriptionModelV2 {
         : new Blob([convertBase64ToUint8Array(audio)]);
 
     formData.append('model', this.modelId);
-    formData.append('file', new File([blob], 'audio', { type: mediaType }));
+    const fileExtension = mediaTypeToExtension(mediaType);
+    formData.append(
+      'file',
+      new File([blob], 'audio', { type: mediaType }),
+      `audio.${fileExtension}`,
+    );
 
     // Add provider-specific options
     if (groqOptions) {
@@ -134,8 +140,8 @@ export class GroqTranscriptionModel implements TranscriptionModelV2 {
           startSecond: segment.start,
           endSecond: segment.end,
         })) ?? [],
-      language: response.language,
-      durationInSeconds: response.duration,
+      language: response.language ?? undefined,
+      durationInSeconds: response.duration ?? undefined,
       warnings,
       response: {
         timestamp: currentDate,
@@ -148,25 +154,28 @@ export class GroqTranscriptionModel implements TranscriptionModelV2 {
 }
 
 const groqTranscriptionResponseSchema = z.object({
-  task: z.string(),
-  language: z.string(),
-  duration: z.number(),
   text: z.string(),
-  segments: z.array(
-    z.object({
-      id: z.number(),
-      seek: z.number(),
-      start: z.number(),
-      end: z.number(),
-      text: z.string(),
-      tokens: z.array(z.number()),
-      temperature: z.number(),
-      avg_logprob: z.number(),
-      compression_ratio: z.number(),
-      no_speech_prob: z.number(),
-    }),
-  ),
   x_groq: z.object({
     id: z.string(),
   }),
+  // additional properties are returned when `response_format: 'verbose_json'` is
+  task: z.string().nullish(),
+  language: z.string().nullish(),
+  duration: z.number().nullish(),
+  segments: z
+    .array(
+      z.object({
+        id: z.number(),
+        seek: z.number(),
+        start: z.number(),
+        end: z.number(),
+        text: z.string(),
+        tokens: z.array(z.number()),
+        temperature: z.number(),
+        avg_logprob: z.number(),
+        compression_ratio: z.number(),
+        no_speech_prob: z.number(),
+      }),
+    )
+    .nullish(),
 });

packages/provider-utils/src/index.ts

@@ -12,6 +12,7 @@ export { isUrlSupported } from './is-url-supported';
 export * from './load-api-key';
 export { loadOptionalSetting } from './load-optional-setting';
 export { loadSetting } from './load-setting';
+export { mediaTypeToExtension } from './media-type-to-extension';
 export * from './parse-json';
 export { parseJsonEventStream } from './parse-json-event-stream';
 export { parseProviderOptions } from './parse-provider-options';

packages/provider-utils/src/media-type-to-extension.test.ts

@@ -0,0 +1,26 @@
+import { describe, it, expect } from 'vitest';
+import { mediaTypeToExtension } from './media-type-to-extension';
+
+describe('mediaTypeToExtension()', () => {
+  it.each([
+    // most common
+    ['audio/mpeg', 'mp3'],
+    ['audio/mp3', 'mp3'],
+    ['audio/wav', 'wav'],
+    ['audio/x-wav', 'wav'],
+    ['audio/webm', 'webm'],
+    ['audio/ogg', 'ogg'],
+    ['audio/opus', 'ogg'],
+    ['audio/mp4', 'm4a'],
+    ['audio/x-m4a', 'm4a'],
+    ['audio/flac', 'flac'],
+    ['audio/aac', 'aac'],
+    // upper case
+    ['AUDIO/MPEG', 'mp3'],
+    ['AUDIO/MP3', 'mp3'],
+    // invalid
+    ['nope', ''],
+  ])('should map %s to %s', (mediaType, expectedExtension) => {
+    expect(mediaTypeToExtension(mediaType)).toBe(expectedExtension);
+  });
+});

packages/provider-utils/src/media-type-to-extension.ts

@@ -0,0 +1,22 @@
+/**
+ * Maps a media type to its corresponding file extension.
+ * It was originally introduced to set a filename for audio file uploads
+ * in https://github.com/vercel/ai/pull/8159.
+ *
+ * @param mediaType The media type to map.
+ * @returns The corresponding file extension
+ * @see https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/MIME_types/Common_types
+ */
+export function mediaTypeToExtension(mediaType: string) {
+  const [_type, subtype = ''] = mediaType.toLowerCase().split('/');
+
+  return (
+    {
+      mpeg: 'mp3',
+      'x-wav': 'wav',
+      opus: 'ogg',
+      mp4: 'm4a',
+      'x-m4a': 'm4a',
+    }[subtype] ?? subtype
+  );
+}