Turn a Morse signal back into text. Pick the decoder that matches what you are holding: a live sound, a recording, or a blinking light.
Decoding Morse is not one problem but three, and the difference is the input. A tone coming out of a speaker, a file sitting on your disk, and a torch blinking in a video need different detection, so MorseKit ships a separate tool for each instead of one that half works on all of them. Lights come in two forms as well, a recording or a lamp in front of you, and the two audio inputs also come combined in a single screen, for when you do not yet know which one you will have.
All of them run entirely in the browser, and they differ in what they measure. Live microphone input is filtered to a narrow band around 600 Hz, so a steady tone is picked out of room noise. A dropped audio file is read straight off the waveform instead: the decoder thresholds amplitude, which means it does not care about pitch but does need the recording to be reasonably loud and clean. The video and camera decoders work on brightness inside a region you select. All of them then derive the dot and dash boundary from the run lengths they observe, so they follow the sender's speed rather than assuming a fixed WPM. Nothing is uploaded to a server in any of them.
Input: Live sound
Point your microphone at a speaker, a key, a buzzer or a radio and read the text as the tones arrive. It listens around 600 Hz, so a steady tone carries over room noise. Nothing is uploaded: the decoding runs in your browser.
Best for: A signal playing right now in the room.
Input: MP3, WAV, OGG
Drop in a recording and get the message back. It reads the waveform rather than the pitch, so any tone works as long as it is clear. Any format your browser can play is accepted.
Best for: A recording you already have on disk.
Input: Flashing light
Select the part of the frame where the light blinks and the decoder reads the flashes. Built for torch signals, lamps and blinking LEDs rather than sound.
Best for: A video of a light, not a tone.
Input: Live light
Point your phone's camera at a blinking light and read the message as it flashes. Tap the preview to put the target on the light; the decoder recalibrates as more of the signal arrives. Works on lamps, LEDs and lights inside games shown on a screen.
Best for: A light blinking in front of you right now.
Input: Mic or file
The microphone and file decoders in one screen, with a switch between them. Use this if you are not sure which input you will end up with.
Best for: Trying both without changing pages.
Available on iOS and Android. Free to download.
Free · iOS & Android · 27 Languages
See how it worksIt depends on what you are holding. If the signal is audible right now, use the Microphone Decoder. If it is a saved recording, use the Audio File Decoder. If it is a blinking light, use the Camera Decoder when the light is in front of you and the Video Decoder when you have a recording of it, because a flashing light carries no tone for an audio decoder to find. The combined Audio Decoder covers the first two in one screen.
No. Every decoder runs entirely in your browser using the Web Audio API and canvas. The file or camera feed never leaves your device, and the microphone or camera permission is the only thing you are asked for.
It depends which one. The microphone decoder listens around 600 Hz, so a tone far off that pitch, or heavy room noise, will defeat it. The audio file decoder ignores pitch entirely and thresholds the waveform, so there the usual cause is a quiet or noisy recording rather than the wrong note: normalise the file and try again. For the video and camera decoders, make sure the selected region covers only the light and that the light is clearly brighter than its background.
Usually yes, within limits. The decoders work out the dot and dash boundary from the run lengths they observe, so they adapt to a steady hand. Very uneven manual sending, where the same operator's dashes vary widely in length, is the case that breaks it.
The MorseKit app decodes offline on iOS and Android, which is the case these web tools cannot cover. The web decoders need the page loaded, but once it is open the decoding itself is local.
