docs: add proposition to karaoke exercise

lal-api-design.md

@@ -99,4 +99,39 @@ Before providing the API specs for the second type of the exercises page, there'
 **Problem:** Flexibility in handling audio.  
 **Option:** Provide both a file path and Base64-encoded audio in the JSON.
 
----
+---
+
+### **Proposition of the Karaoke Exercise JSON Structure**
+
+This proposition uses one solution of the different design challenges mentioned above and is designed for a karaoke type exercise where audio is associated with tasks that guide the user through the exercise. The **audio** path points to the MP3 file used in the exercise. Each task contains a segment of the lyrics that should be highlighted in sync with the audio playback. 
+
+In this structure:
+- The **audio** is stored as a file path to the MP3.
+- The **tasks** are broken down by segments of the audio, and each segment corresponds to a specific time range defined by **start_time** and **end_time**.
+- The **highlighted_text** field will now represent **words** rather than full sentences, providing more granular control for highlighting.
+
+Here’s the updated JSON example:
+
+```json
+{
+    "exercise_id": 1,
+    "title": "Karaoke Exercise",
+    "audio": "/path/to/audio/file/karaoke_exercise.mp3",
+    "tasks": [
+        {
+            "id": 1,
+            "text": "The lyrics of the audio content subtask",
+            "start_time": 0,
+            "end_time": 30,
+            "highlighted_text": "The first word of the first sentence"
+        },
+        {
+            "id": 2,
+            "text": "Continue with the next part of the lyrics",
+            "start_time": 31,
+            "end_time": 60,
+            "highlighted_text": "The first line of the second part of the audio"
+        }
+    ]
+}
+```