Local by design
The signal stays in memory on your device. Stopping or leaving the page closes the microphone stream and audio context.

Sing or play one clear note and get stable feedback. Detect the note, frequency, and tuning in cents with voice-aware octave correction, right in your browser.
Play or sing a steady note
Private, on-device audio
Allow browser access · Hold one clear note
Voice range
Singers can select a range to reduce octave errors. Leave unrestricted for instruments.
A single microphone signal becomes three complementary readings, so you can identify the note and see how accurately it is tuned.
YIN / LOCAL DSP
PitchDetect measures repeating patterns in the microphone signal with the YIN method, converts the frequency to the nearest equal-tempered note, and reports the difference in cents.
The signal stays in memory on your device. Stopping or leaving the page closes the microphone stream and audio context.
A noise gate, confidence threshold, and multi-frame smoothing prevent weak or unstable sounds from appearing as certain notes.
Use it with guitar, bass, strings, brass, woodwinds, piano tones, humming, or a sustained singing voice.
Answers about pitch, tuning, privacy, and troubleshooting.
A pitch detector estimates the fundamental frequency of a sound; a note detector maps that frequency to the nearest musical note. PitchDetect shows the frequency, note, and tuning difference in cents together.
It sets the A above middle C to 440 cycles per second. Other equal-tempered notes are calculated from that reference. You can adjust the reference from 420 to 460 Hz.
A cent is one hundredth of a semitone. Negative cents are flat, positive cents are sharp, and values close to zero are in tune with the selected reference.
Yes. Analysis runs locally in your browser. PitchDetect does not record, upload, or persist microphone audio, and Stop closes the active media tracks.
Check microphone permission and the selected input, move closer, reduce background noise, and hold one steady note. Weak or unstable input is intentionally withheld. Singers can select their voice range if the octave appears wrong.
Yes. It works best with a clear, sustained, mostly monophonic sound. Chords, heavy effects, room echo, and background music can make the fundamental pitch ambiguous.