Using Mullo
Recording
Phone, watch, or a typed note.
On the phone
- Start recording on the Record screen. A level meter shows the microphone is hearing you. Pause, resume, or discard (tap twice, so it can't happen by accident).
- Summarise as: choose the format for this recording before you start, or leave the default from Settings.
- Live preview (Settings, Recording and transcription) shows text as you talk, made by your phone. The final transcript is written by the AI when you stop. Turn it off and the audio goes only to the AI.
- Lock the phone if you like: in the iPhone and Android apps the microphone keeps going with the screen off. On Android a "Mullo is recording" notification shows while it records; tap it to come back. In a phone's browser, keep Mullo on screen while recording.
- Long meetings are fine, up to about 3 hours in one recording. The phone sends the recording and collects the result later, so a weak signal or a closed app doesn't lose anything.
On a Wear OS watch
One press starts, the next stops and sends. The recording moves to your phone by itself and is written down there.
- One buzz = recording now: the microphone is on, start talking. Two buzzes = stopped and saved. One long buzz = it couldn't start; open Mullo on the watch to see why.
- Double-press the side button (Samsung Galaxy Watch): watch Settings, Advanced features, Customize buttons (or keys), Double press, Mullo record.
- A button on your watch face (any Wear OS watch, including ones whose button can't open an app): press and hold the watch face, Edit (or Customize), choose a spot, then Mullo record. Tap it to start; it turns into a stop button while recording.
- The very first press after the watch has been idle for a while can take a moment longer: the watch is opening Mullo. Wait for the buzz before you talk.
On an Apple Watch
Open Mullo on the watch, or add its button to your watch face (press and hold the face, Edit, choose a spot, Mullo), and tap to record. Tap again to stop. The note goes to the Mullo app on your iPhone and is written down there.
- If the iPhone is out of reach, the watch keeps the note and sends it as soon as the two are connected again, even if Mullo is closed on the iPhone.
- A note stays on the watch until the iPhone has it, so nothing is lost if the watch app closes mid-send.
Pause, and calls, on a watch
- Pause: while recording, the small button beside Stop pauses. The microphone goes off and the timer stops. Tap it again (Play) to go on in the same recording. A pause you make never ends by itself.
- A call on the watch: one that rings and isn't answered doesn't stop the recording. Answer it on the watch (or Siri or another app takes the microphone) and the recording pauses, then goes on by itself when the call ends. If the watch can't go on by itself, it buzzes and shows Paused: tap Play or Stop. What was recorded before is always kept.
- The note shows each pause at its place in the transcript ("Paused 6 s, call"), and the summary is told a piece is missing there, so it doesn't join what was said before and after.
- Pause the watch before you take a call on your phone, or it may record your side of the call.
How long a recording can be: Limits and AI models.
A typed note
Under the record button, Or type a note: write it and Save note. It gets a summary and tasks like a recording.
Languages
Mullo writes down about 110 languages. About 60 of them go to Mullo's main speech model, which labels who said what and follows several languages in one note: Afrikaans, Albanian, Arabic, Azerbaijani, Basque, Belarusian, Bengali, Bosnian, Bulgarian, Catalan, Chinese, Croatian, Czech, Danish, Dutch, English, Estonian, Filipino, Finnish, French, Galician, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Marathi, Norwegian, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Vietnamese, Welsh. The other languages are written down by a second AI model, which labels speakers too.
In Settings, Languages, set the language you speak and, if you like, a different language to write the transcript in. A transcript in another language is always written by the second model.
- Auto (mixed languages) listens for any language, and for more than one in a note, such as Hebrew with English words. If the main model hears a language it doesn't know, or isn't sure what it heard, the second model writes the note instead.
- Pick a specific language when you nearly always speak the same one, and above all when it isn't on the list above: the note goes straight to the model that knows it. Re-transcribe with a language picked keeps the transcript to that language.