Suno API Aligned Lyrics: Word Timestamps and LRC

Short answer: call POST /api/lyrics/timeline with the song ID and model. It returns an array of timed words with word, start_s, end_s, success and alignment confidence. Use those values for LRC, karaoke highlighting, short-video captions or a synchronized player.

Plain lyrics versus timed lyrics

Request the timeline

curl -X POST "https://www.suno-api.io/api/lyrics/timeline" \
  -H "Authorization: Bearer $SUNO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "suno-v6",
    "song_id": "YOUR_SONG_ID"
  }'

[
  {
    "word": "city lights",
    "start_s": 1.12,
    "end_s": 1.58,
    "success": true,
    "p_align": 0.99
  }
]

The response is an array, without an extra data envelope.

Generate LRC

def to_lrc(items):
    lines = []
    for item in items:
        text = (item.get("word") or "").strip()
        if not text or not item.get("success", True):
            continue
        start = max(0.0, float(item.get("start_s") or 0))
        cs = int(start * 100 + 0.5)
        minute = cs // 6000
        second = (cs % 6000) // 100
        centi = cs % 100
        lines.append(f"[{minute:02d}:{second:02d}.{centi:02d}]{text}")
    return "\n".join(lines)

Karaoke highlighting

1. Read current playback time.
2. Find the item where start_s <= time < end_s.
3. Mark previous items as sung and the current item as active.
4. Recompute by time after any seek; do not rely only on array position.
5. Keep low-confidence items available for manual correction.

FAQ

Is timeline output the same as normal lyrics?

No. It contains start and end timing for rendering synchronized text.

Is a prebuilt LRC available for every song?

No. Generate LRC from timeline data when a direct LRC download is unavailable.

Can I store the timeline permanently?

Yes. Store the data and song ID; a confidence-aware preview is still recommended.

Use the timeline API reference, MIDI guide, and media download API.

← Back to all articles