Your taps are the only tempo input
YouTubeBPMFinder.online is a local companion for a video you control elsewhere. Start playback in another window or on another device, choose a recurring pulse, and tap the large surface with Space, Enter, a mouse, or touch. The page times those human actions. It does not receive the video, hear audio, or know what you are watching.
That separation is the product promise. There is no YouTube URL box, embedded player, authorization prompt, microphone permission, upload, or automatic tempo claim. The observer supplies a pulse through taps, and the browser exposes how those intervals behave.
One tap creates a point; two create an interval
The first tap starts a session. Each later tap closes another interval. Eight taps create seven intervals. The page uses performance.now(), a monotonic clock available to the current document, so calculations depend on elapsed page time rather than a readable calendar timestamp.
After at least four taps, a provisional result can appear. A longer clean run usually offers better context, but tapping through a section change can make the summary less representative. Choose a recent-window length from 4 through 32 intervals and stop when the musical evidence changes.
Several summaries reveal different features
Overall observed BPM divides every completed interval by total elapsed time: 60,000 × interval count ÷ elapsed milliseconds. Median-interval BPM uses the median interval duration, which resists one extreme gap better than a simple average. When at least ten intervals exist, a 20% trimmed result removes 10% from each end of the sorted interval list before averaging. The recent-window result uses only the latest chosen intervals.
The stability panel shows the slowest and fastest interval-equivalent BPM in that recent window and their range. It is not a grade. Human anticipation, a missed tap, swing, a fill, browser scheduling, or real performance movement can widen the result.
A twelve-tap example
Imagine twelve taps producing eleven intervals around 500 milliseconds. Total elapsed time is 5,500 milliseconds, so overall observed BPM is 60,000 × 11 ÷ 5,500 = 120. If the median interval is 498 milliseconds, median-interval BPM is about 120.48. Small differences among summaries reveal the tap pattern; they do not automatically identify an error.
If external playback was at 0.75×, the page corrects the overall observed 120 by division: 120 ÷ 0.75 = 160 source BPM. Half and double source candidates become 80 and 320. The card must retain the 0.75× setting so 160 is not mistaken for the rate actually experienced while tapping.
Select the pulse before selecting a result
A hi-hat subdivision can invite taps at twice the practical beat. A backbeat can invite half-time taps. The half, base, and double candidates are a listening aid, not a detector. Choose the pulse that matches your task and label it briefly, such as “kick pulse” or “main groove.” Avoid personal or confidential text because labels are unnecessary for arithmetic and can enter the clipboard.
Guides cover subdivision traps, live recordings, long mixes, speed correction, and observation notes. Narrow tools plan a tap duration, audit one completed card, compare card methods, inspect interval spread, show pulse-level candidates, or flag possible outliers. Their inputs remain separate so a visitor can select the exact task without confusing one result for another.
Product limitations
Tap timing cannot identify a song, infer meter, isolate drums, distinguish intentional swing from error, or establish the tempo of every section. A session describes one observer, one pulse choice, one playback speed, and one period of tapping. Browser clock resolution and event scheduling can introduce small variation, especially when the page is backgrounded or the device is busy.
The site never fetches, embeds, downloads, records, parses, or listens to YouTube media. It does not call the YouTube API, use a microphone, read a player setting, access an account, search a song library, store a history, or send tap values to an analytics vendor.
Local memory and independent identity
Tap times, optional labels, settings, and results live in current-page memory. Reset removes the session, and closing or reloading ends the defined state. Copy places a plain-text card on the device clipboard when browser permission allows; the operator does not receive it.
YouTube is a trademark of Google LLC. This independent site is not affiliated with, endorsed by, sponsored by, or operated by Google LLC or YouTube. Its violet, amber, aqua, and lavender pulse-lantern identity avoids logos, thumbnails, player controls, and red-and-white trade dress.
Frequently asked questions
Can I paste a YouTube link?
No. Control playback elsewhere and tap on this page. No URL retrieval or embed exists.
Which key should I use?
Space or Enter works when focus is on the tap surface. Pointer and touch input are also supported. Held-key repeats are ignored.
Is median BPM the correct answer?
No summary is automatically musically correct. Median interval timing is simply less influenced by one extreme interval.
Does slowing playback change source BPM?
It changes observed tap rate. Divide observed BPM by the speed factor to estimate the source-time rate.
Tap a clean run and inspect the constellation
Open the local tap surface, choose one pulse, and collect at least eight taps. Compare overall, median, trimmed, and recent summaries before copying a card with the speed and pulse label.