How to Scrape TikTok Sound & Music Videos in 2026

If you want to scrape TikTok sound and music videos — every clip riding a given track, with the engagement and creator behind each one — you are working with one of the most valuable and most overlooked signals on the platform. On TikTok, the sound is the unit of virality. A single audio clip can carry a song from zero to the charts in a week, and the only way to see that spread is to pull the full catalogue of videos using it. This guide covers why videos-by-sound data matters, how to feed a sound into the scraper, what comes back, and what music teams, trend spotters and creators actually do with it.

Why videos-by-sound is the signal that matters

Most TikTok scraping is organized around hashtags, keywords or accounts. That’s fine for surface-level monitoring, but it misses how TikTok actually distributes content. A song doesn’t break out because of a hashtag — it breaks out because thousands of creators independently attach the same sound to their videos. The sound is the carrier. If you want to know whether a track is gaining traction, you don’t count hashtags; you count the videos using that specific audio and watch how fast that count climbs.

That makes a sound-centric scrape fundamentally different in value:

• Audio trend tracking. Run the same sound on a schedule and the video count plus aggregate engagement gives you a clean growth curve. You see a track climbing days before it shows up on any chart.
• Music marketing attribution. When a label or artist seeds a track, this is how you measure whether the seeding worked — who picked it up, how many views the early videos pulled, and which creators drove the spread.
• Creator discovery from the inside. The creators powering a sound are, by definition, the ones whose audience responds to that style of audio. They are the people you want to brief next.

A single viral sound can have thousands of videos, which is exactly why this is a bulk-export problem and not something you can eyeball in the app.

Why scraping by sound is clean

The TikTok music page — the one you reach by tapping the spinning sound disc on any video — is public. There is no login wall on it, and there’s no need to attach your own account to read it. The scraper opens that public page over a residential proxy and intercepts the video feed as it paginates.

That posture matters:

• No login, no cookies. You never touch credentials, so there’s no account to get restricted or banned. This is public video data only.
• Residential proxies handle the rest. TikTok requires residential egress to serve the feed reliably; the actor ships pre-configured for that, so you don’t assemble a proxy stack yourself.
• Overflow-safe engagement. View and like totals on big videos are read from TikTok’s overflow-safe stats, so a 4.6M-view clip comes back as 4600000, not a truncated “4.6M” string you’d have to reparse.

Input: a sound URL or a music id

The scraper takes one required thing — a list of sounds — plus a few optional caps.

Field	Type	Description
sounds	array	Required. Music page URLs or music ids.
maxVideosPerSound	integer	How many videos to collect per sound. Default 200.
maxResults	integer	Global cap on videos saved. 0 = unlimited.
proxyConfiguration	object	Residential proxy (required by TikTok). Pre-configured for you.

You can pass each sound as a full music page URL or as just its numeric music id — both resolve to the same place. Mixing the two formats in one run is fine.

{
  "sounds": [
    "https://www.tiktok.com/music/original-sound-7608483104361122593",
    "7608483104361122593"
  ],
  "maxVideosPerSound": 300
}

How to find a sound’s URL or id

Open any TikTok video and tap the spinning sound disc in the bottom-right corner — that opens the music page. The page URL ends with the numeric music id, e.g. .../music/original-sound-7608483104361122593. Paste the whole URL, or just the trailing id. Both work.

Pagination and caps

A sound’s feed paginates, and the actor walks that pagination automatically, collecting videos until it hits your maxVideosPerSound limit (default 200). Raise that number to go deeper into a high-volume sound. TikTok’s own reported total-video count for the sound is surfaced in the logs and in each record’s metadata, so you can see how much of the catalogue you captured versus the full universe.

Two caps work together:

• maxVideosPerSound bounds each individual sound — useful when you pass several sounds and want a balanced sample from each.
• maxResults is a global ceiling across the whole run. Set it to 0 for unlimited, or use it to keep a multi-sound run inside a predictable budget.

Across multiple sounds in one run, results are de-duplicated, so a video that happens to surface under more than one sound is saved once.

Output: exact fields per video

Each record is one video using the sound, carrying its engagement, its author, and the sound’s own metadata stamped onto it:

{
  "videoId": "7600000000000000000",
  "url": "https://www.tiktok.com/@amelietjer/video/7600000000000000000",
  "description": "#yachting #superyacht #stewardess",
  "postedAt": "2026-02-18T14:05:00.000Z",
  "durationSec": 18,
  "playCount": 4600000,
  "likeCount": 211600,
  "commentCount": 1900,
  "shareCount": 7400,
  "saveCount": 22300,
  "hashtags": ["yachting", "superyacht", "stewardess"],
  "author": { "username": "amelietjer", "nickname": "Amélie", "verified": false },
  "soundId": "7608483104361122593",
  "soundTitle": "original sound",
  "soundAuthor": "Way Of House",
  "soundUrl": "https://www.tiktok.com/music/original-sound-7608483104361122593",
  "scrapedAt": "2026-06-03T18:00:00.000Z"
}

Field by field:

• videoId — stable per-video identifier. Use it as your join/dedup key across scheduled runs.
• url — direct link to the video.
• description — the caption, hashtags inline.
• postedAt — ISO timestamp of when the video was published. This is what you bucket by to build a growth curve.
• durationSec — video length in seconds.
• playCount, likeCount, commentCount, shareCount, saveCount — the full engagement set, as accurate integers.
• hashtags — parsed array, no #.
• author — username, nickname and verified flag for the creator.
• soundId, soundTitle, soundAuthor, soundUrl — the sound’s own metadata (title, artist, page URL) stamped on every video, so a multi-sound export stays self-describing.
• scrapedAt — capture timestamp. Record it: a sound’s feed is a moving target, so you need to know when each row was taken.

Export as JSON, CSV, Excel or HTML, or pull straight from the Apify API and webhooks.

▶ Run the TikTok Sound & Music Video Scraper — one sound in, every video using it out, with engagement, creators and sound metadata. No login, no cookies, no ban risk. $5 per 1,000 videos.

Schema choices for tracking over time

If you’re storing this for trend analysis rather than a one-off pull, a few decisions pay off:

• Key on videoId, partition by soundId. That gives you a clean “videos per sound over time” table for free.
• Always keep postedAt and scrapedAt separate. postedAt tells you when the video was made; scrapedAt tells you when you observed its stats. Engagement on a video keeps climbing, so two scrapes of the same videoId at different scrapedAt values is exactly how you measure velocity.
• Snapshot the engagement fields per run, don’t overwrite. The growth of playCount on individual videos between runs is the most actionable signal you have — a track where existing videos keep accumulating views is a track that’s still alive.
• Aggregate at the soundId level for the curve. Count of videos + sum of playCount per scrapedAt is your headline trend line.

Use cases

What teams actually do with videos-by-sound data:

• Music labels and A&R. Seed a track, then run its sound daily to see who picks it up and how fast. Catch a song gaining organic traction before it charts — that’s the entire edge in A&R. The author and verified fields turn the spread into a shortlist of creators to sign or partner with.
• Music marketing and promotion. Measure whether a paid creator campaign actually moved the sound. The video count and aggregate views attributable to a track are your proof-of-traction for a campaign report.
• Trend spotters and agencies. Track a basket of rising sounds and surface the ones whose growth curve is steepening. A scheduled, structured feed across a set of sounds is a sellable trend-intelligence product; a manual scroll through the app is not.
• Creators picking their next sound. Before you build a video around an audio clip, pull its feed: is the sound still climbing or already saturated? Which formats and niches are winning with it? description, hashtags and engagement across the existing videos tell you what’s working before you commit.
• Rights and royalties research. Audit how widely a track is being used across TikTok — useful for sync, licensing and usage disputes where you need a defensible count of videos using a given sound.

The common thread is the same sounds, refreshed over time. A single snapshot tells you a sound is popular. A scheduled feed tells you whether it’s accelerating or fading — and that’s the part worth paying for.

Cost math

Pricing is $5 per 1,000 videos plus a trivial per-run start fee. The cost is per result, so you only pay for videos you actually collect.

Concrete example: tracking 5 sounds, capturing up to 200 videos each, refreshed weekly, is roughly 4,000 videos a month — about $20/month at $5 per 1,000 videos for a continuously refreshed, multi-sound audio-trend feed.

Compare that to building it yourself: the public music page is reachable, but reliably intercepting the paginated video feed over residential proxies, parsing overflow-safe engagement counts, de-duplicating across sounds, and keeping all of it working as TikTok shifts its internals is exactly the maintenance the managed actor absorbs.

FAQ

What counts as a “sound”? Any TikTok music page — original sounds, licensed songs or remixes. Pass its full music page URL or just the numeric music id.

Will I get every video using the sound? You’ll get up to maxVideosPerSound (default 200); raise it to go deeper. TikTok’s own reported total for the sound is included in the logs and metadata, so you always know what fraction of the full catalogue you captured.

Can I track several sounds in one run? Yes. Add multiple URLs or ids to the sounds array. Results are de-duplicated across them, and each video carries its soundId/soundTitle so you can split them apart downstream.

Do I need a login or cookies? No. This reads public video data via residential proxies only — there’s no account involved and no ban risk.

How do I track a sound’s growth over time? Schedule the same sound on a recurring run and store each pull with its scrapedAt. Comparing video count and aggregate playCount across runs gives you the growth curve.

What’s the best way to find creators driving a sound? Sort the output by playCount and look at the author field — the accounts behind the top-performing videos are the ones powering the sound’s spread, and the verified flag helps you separate established creators from breakout newcomers.

Wrapping up

On TikTok, the sound is the engine of virality, and videos-by-sound is the cleanest way to see a track move. Because the music page is public, you can scrape TikTok sound and music videos with no login, no cookies and no account risk — just a music URL or id in, and every video riding that sound back out, complete with engagement, creators and the sound’s metadata. Run it once for a snapshot, or schedule it across a basket of sounds to watch audio trends climb in real time.

▶ Open the TikTok Sound & Music Video Scraper on Apify — track a song’s spread across TikTok, pull thousands of videos per sound, export to JSON/CSV/Excel. $5 per 1,000 videos. Start with Apify’s free monthly credit.

Filed under: how to scrape TikTok sound & music videos.