diff --git a/packages/api/src/utils/textToSpeech.ts b/packages/api/src/utils/textToSpeech.ts
index c348b9c4b..d9cf0af7e 100644
--- a/packages/api/src/utils/textToSpeech.ts
+++ b/packages/api/src/utils/textToSpeech.ts
@@ -61,7 +61,7 @@ export const createSpeechMarks = async (
   }
   try {
     const data = await client.synthesizeSpeech(params).promise()
-    return data.AudioStream as string
+    return (data.AudioStream as Buffer).toString()
   } catch (error) {
     logger.error('Unable to create speech marks', { error })
     throw error
@@ -90,7 +90,7 @@ export const createAudioWithSpeechMarks = async (
       speechMarks,
     }
   } catch (error) {
-    logger.error('Unable to create audio with speech marks', { error })
+    logger.error('Unable to create audio with speech marks', error)
     throw error
   }
 }
diff --git a/packages/api/test/utils/textToSpeech.test.ts b/packages/api/test/utils/textToSpeech.test.ts
new file mode 100644
index 000000000..62a3e53c2
--- /dev/null
+++ b/packages/api/test/utils/textToSpeech.test.ts
@@ -0,0 +1,23 @@
+import 'mocha'
+import {
+  createAudioWithSpeechMarks,
+  TextToSpeechInput,
+} from '../../src/utils/textToSpeech'
+import { expect } from 'chai'
+import { generateFakeUuid } from '../util'
+
+describe('textToSpeech', () => {
+  describe('createAudioWithSpeechMarks', () => {
+    it('should create an audio file with speech marks', async () => {
+      const input: TextToSpeechInput = {
+        id: generateFakeUuid(),
+        title: 'Hello World',
+        text: 'Hello World',
+        engine: 'standard',
+      }
+      const output = await createAudioWithSpeechMarks(input)
+      expect(output.audioUrl).to.be.a('string')
+      expect(output.speechMarks).to.be.a('string')
+    })
+  })
+})