TikTok Sound Scraper & Monitor - Video Count & Audio Growth

Track TikTok sounds from their URLs: how many videos use a sound, its title, author, length, album and original-sound flag, the artist shown on TikTok's sound card, and the top videos using it with their view counts. The video count comes with the step TikTok rounded it to, so you know that 1300000 is TikTok's "1.3M", not an exact number.

Turn on monitoring mode, schedule the Actor, and a run returns a sound only when its video count changed since the row last returned for it, with the change and the videos per hour. Sounds TikTok's sound card does not return come back as free rows with TikTok's own error code.

What you can use it for

• Trend tracking: schedule monitoring mode on the sounds you care about and see which ones are gaining videos, and how fast (videoCountDelta, videosPerHour).
• Music marketing and campaign reporting: paste the sounds of a release or campaign and export video counts, top videos and their views to a spreadsheet or dashboard.
• Creator and brand research: see which videos TikTok shows first for a sound, and who made them.
• Sound catalogues: get the title, author, length, album and TikTok's original-sound flag for a list of sound ids.

How it works

1. Paste sound URLs (https://www.tiktok.com/music/SugarCrash-6920125567752734722), embed URLs (https://www.tiktok.com/embed/music/6920125567752734722) or bare numeric sound ids.
2. For each sound the Actor opens TikTok's public sound card (https://www.tiktok.com/embed/music/<id>) and checks that TikTok answered with the same sound id you asked for. The card gives the video count, the artist name, the cover image and up to 10 top videos.
3. The card has no title, so the Actor opens the page of one public video from the card and reads the sound's title, author, length, album and original-sound flag from it - only if that video's sound id is the sound you asked for (from Apify on 2026-09-15, 19 of 19 videos opened this way carried the sound's own id). If the first video does not work it tries a second one; if neither does, the row is still returned with the counts and soundDetailsStatus says why the title is missing.
4. Sounds the card does not return get a free row with TikTok's error code.

TikTok's own sound API (/api/music/detail) needs a signed request: it answered with an empty response (HTTP 200, 0 bytes) from Apify and from a home connection alike on 2026-09-15, so this Actor does not use it.

Input

Field	Default	What it does
sounds	example sounds	Sound URLs, embed URLs or numeric sound ids, one per line (commas also separate entries, and so do spaces between entries that are all sound URLs or ids). A repeated sound is read and charged once; the repeat gets a free duplicate row. Short links (vm.tiktok.com) are not followed.
maxResults	100	With monitoring off, stop after this many rows. In monitoring mode it does not cut changes. Up to 1,000 entries per run.
includeSoundDetails	true	Open one video page per returned sound to read the title, author, length, album and original-sound flag. Off = one request per sound and those columns stay empty. The price is the same.
monitoringMode	false	Remember each sound's video count and return a sound only when it changes.
resetMonitoringState	false	Forget every remembered count and start a fresh baseline. Turn it off again after one run.
useProxy	true	Retry a page through a US proxy only if TikTok answers it with a bot check.

{
    "sounds": [
        "https://www.tiktok.com/music/SugarCrash-6920125567752734722",
        "https://www.tiktok.com/music/Come-and-Get-Your-Love-6688272830112566021"
    ],
    "monitoringMode": true
}

If you leave sounds out of the input with monitoring off, the two example sounds above are read (their URLs appear in inputSound and the run log says so). In monitoring mode sounds is required, so a schedule never pays to watch the examples. An empty list is rejected with a free invalid-input row instead of reading the examples.

Output

One row per sound, shortened (2 of the 10 top videos shown). The counts and details were read from Apify on 2026-09-15 (UTC); the monitoring fields show what a later run returns after the video count moves from 1.2M to 1.3M:

{
    "status": "ok",
    "inputSound": "https://www.tiktok.com/music/MONTERO-6942551356402042881",
    "soundId": "6942551356402042881",
    "soundUrl": "https://www.tiktok.com/music/MONTERO-Call-Me-By-Your-Name-6942551356402042881",
    "embedUrl": "https://www.tiktok.com/embed/music/6942551356402042881",
    "title": "MONTERO (Call Me By Your Name)",
    "authorName": "Lil Nas X",
    "embedArtist": "dreamboy.",
    "isOriginal": false,
    "durationSeconds": 40,
    "album": "MONTERO (Call Me By Your Name)",
    "soundDetailsStatus": "ok",
    "soundDetailsFromVideoId": "6947415259955088646",
    "videoCount": 1300000,
    "videoCountRoundingUnit": 100000,
    "videoCountIsRounded": true,
    "videoCountAtLeast": 1200001,
    "videoCountAtMost": 1399999,
    "tiktokRegion": "US",
    "tiktokDataCenter": "US-TTP2",
    "topVideos": [
        {
            "videoId": "6947415259955088646",
            "videoUrl": "https://www.tiktok.com/@jorgechaconh/video/6947415259955088646",
            "authorUsername": "jorgechaconh",
            "description": "You know that you can #callmebyyourname 👼 @lilnasx 🖤",
            "playCount": 87800000,
            "playCountRoundingUnit": 100000,
            "isPrivate": false
        },
        {
            "videoId": "6946346066350492933",
            "videoUrl": "https://www.tiktok.com/@worms.forbrains/video/6946346066350492933",
            "authorUsername": "worms.forbrains",
            "description": "well I don’t have a pole, but I have a skateboard 😈 @lilnasx #cmbyn",
            "playCount": 31400000,
            "playCountRoundingUnit": 100000,
            "isPrivate": false
        }
    ],
    "topVideosCount": 10,
    "change": "count-changed",
    "isFirstCheck": false,
    "previousCheckedAt": "2026-09-15T13:36:00.000Z",
    "previousCountReadAt": "2026-09-14T14:36:00.000Z",
    "previousVideoCount": 1200000,
    "previousTiktokRegion": "US",
    "hoursSincePreviousCount": 24,
    "videoCountDelta": 100000,
    "videoCountDeltaAtLeast": -99998,
    "videoCountDeltaAtMost": 299998,
    "videosPerHour": 4166.67
}

Every row also has source, scrapedAt and coverUrl (the sound's cover image as TikTok links it; 7 of the 11 cover links read from Apify on 2026-09-15 carried an expiry time (x-expires) about two days later, so save the image rather than the link).

The video count is TikTok's rounded figure

TikTok's sound card shows the video count rounded. Measured on 2026-09-15 across 39 sounds read from a home connection and 11 read from Apify:

Video count	Rounded to	Example
below 10,000	not rounded	647
10,000 to 999,999	steps of 100	952000
1,000,000 and above	steps of 100,000	1300000

videoCountRoundingUnit is that step, and videoCountAtLeast / videoCountAtMost are the lowest and highest whole numbers the shown figure could stand for, taking one step either way because the direction of TikTok's rounding has not been measured. In monitoring mode videoCountDeltaAtLeast / videoCountDeltaAtMost give the same range for the change, which is why a shown jump of +100000 can be anything from -99998 to +299998. A sound can gain videos without its shown count changing, so a large sound is only returned when its count crosses a rounding step.

This is the same figure TikTok's own sound page shows: for 2 of 2 sounds opened in a browser from a home connection on 2026-09-15, the page said "98.8K videos" and "3.9M videos" and the card read from the same connection right after gave 98800 and 3900000.

Reading the same sound again from the same place gave the same figure: 13 of 13 sounds showed the same count on four reads from the same connection within about 30 minutes, and reading sounds from Apify twice in a row gave matching counts.

Counts depend on which TikTok region answers

TikTok answers from different servers in different countries, and they do not show the same video count. On 2026-09-15 the same 13 sounds were read at the same moment from Apify (tiktokRegion US) and from a connection in Japan (JP): 4 of 13 differed, for example 952000 from Apify against 910800 from Japan, and 647 against 648.

So every row says which region answered (tiktokRegion, tiktokDataCenter). Requests from Apify are answered by TikTok's US servers, and the proxy used after a bot check is located in the US too. In monitoring mode a count is only compared with a count from the same region: if a different region answers, the row comes back as change: "region-changed" without a change figure, and later runs compare against that region.

Top videos

topVideos is the list TikTok puts on the sound card: up to 10 videos, not all videos that use the sound, and not sorted by views. The list can change length between reads (10, 9, then 10 videos for the same sound on 2026-09-15), so it is never used to decide whether a sound changed.

playCount is TikTok's rounded view count. The card stores it as a 32-bit number, which overflows for videos above about 2.1 billion views: Zach King's 2.4-billion-view video is -1894967296 on the card. Such a value is returned as null rather than as a wrong number, and so is any value that cannot be one of TikTok's rounded figures.

Title, author and the artist on the card

title, authorName, durationSeconds, album and isOriginal come from the video page (see How it works). embedArtist is the artist name printed on the sound card, which is not always the same as authorName: 2 of 39 sounds differed on 2026-09-15 (for "MONTERO (Call Me By Your Name)" the card says dreamboy. and the video page says Lil Nas X). Both are returned as TikTok gives them.

isOriginal is TikTok's own original flag, passed through unchanged. It does not always match the title: the sound titled "original sound - NASA" has it set to false.

soundDetailsStatus is ok, not-requested (includeSoundDetails is off), no-public-video (the card lists no public video), different-sound (the video page belongs to another sound), video-unavailable, blocked or unreadable. The row is charged either way, because the video count was read.

Rows that are never charged

Status	Meaning
sound-not-returned	TikTok's sound card answered HTTP 400 with its own error code (tiktokErrorCode). It does this for ids that do not exist (100001), and also for some sounds that do exist: 5 of 44 sounds taken from real video pages got it on 2026-09-15, with codes 10219 and 10218, and 4 of those 5 were not original sounds (licensed or library music). So this row does not claim the sound does not exist or was removed.
different-sound-returned	TikTok answered with a different sound id than the one asked for.
no-count-returned	The card answered without a usable video count.
blocked	TikTok answered with a bot check or an empty response, even after retrying.
unreadable	The card could not be read.
invalid-input	The entry is not a sound URL or id, or it is a short link.
duplicate	The same sound appeared earlier in the list.
no-change	Monitoring mode: no sound checked had a different video count.
not-checked	The end of the list was not read because maxResults rows were returned or more than 1,000 entries were given.
budget-reached	The run's maximum total charge had no room for another row (or check and row).

Pricing

• $4.00 per 1,000 sound rows, charged only for rows with status: "ok".
• In monitoring mode, also $0.30 per 1,000 sound checks: every sound whose video count was read costs a check, whether or not it changed. Example: 50 sounds checked every hour is 36,000 checks a month, which is $10.80, plus $4.00 per 1,000 rows for the sounds that changed.

The Actor reads only as many sounds as the run's maximum total charge can pay for (with monitoring on, a check and a change row each), and a free row says how many were not read.

Monitoring mode

• The first run returns every sound once with change: "first-check".
• Later runs return a sound with change: "count-changed" when the shown video count differs from the row last returned for it, or "region-changed" when a different TikTok region answered.
• A sound that is not returned keeps the count of the row last returned for it, so the next change adds up from there.
• Counts are remembered per sound for this Actor, not per list. Do not put the same sound in two schedules that can run at the same time: each run merges only the sounds it handled into the remembered counts and checks its write, but Apify's key-value store has no atomic update, so two runs finishing at the same moment can still overwrite each other.
• If Apify restarts or moves a run, it continues after the sounds it had already handled, and a sound whose row is already in the run's dataset is not read or charged again (in monitoring mode its count is remembered from that row).

Limits

• Up to 1,000 entries per run, read one at a time with a short pause between requests. Measured from Apify on 2026-09-15: 2.31 seconds per sound for 17 sounds whose title was read, and 1.12 seconds per sound for 41 sounds where only the sound card was read, so 1,000 sounds take roughly 40 minutes with includeSoundDetails on. The default run timeout is 3,600 seconds; raise it for longer lists.
• The Actor does not return the full list of videos using a sound; topVideos is TikTok's card list of up to 10.
• It does not find trending sounds for you; give it the sounds to read or watch.