Skip to content

AI Assistant · Voice notes · New

Say it once. Your WhatsApp assistant does the rest.

Record a voice note in your “Message yourself” chat, in your own language. whatsappx.si transcribes it, shows you what it heard and runs it like a typed command.

NewUpdated
On this page

In short

Record a voice note in your “Message yourself” chat, up to 5 minutes, in any language. whatsappx.si transcribes it, replies with what it heard and runs it like a typed command. Only the words are kept, not the audio. It works on your own WhatsApp, with the AI Assistant switched on.

How does voice-to-task work?

  1. Switch on the AI Assistant

    Link your WhatsApp by QR code and turn the AI Assistant on in your workspace. Voice notes use the same assistant.

  2. Record a voice note to yourself

    Open your “Message yourself” chat on WhatsApp and hold the microphone. Speak your task, for example “ask Priya to fix the banner and remind her in two hours”. A note can be up to 5 minutes.

  3. It shows you what it heard

    The assistant transcribes the note and replies with what it heard (🎤 "...") together with what it did, so you can spot a mistake. If it misheard, pause the assistant or send a correction.

  4. It runs the command

    The transcript is run exactly like a typed command: it messages the person, tracks the task, follows up and tells you when they reply.

Which languages work?

You can speak in your own language. whatsappx.si detects the language automatically, so you do not choose one, and it works in any language the speech provider can transcribe.

It handles these languages best, including speech that mixes them, such as Hinglish:

Handled bestNotes
English, Hindi (and Hinglish)Mixed speech is supported
Spanish, French, German, Italian, Portuguese, DutchDetected automatically
Russian, JapaneseDetected automatically

Other languages are detected and transcribed too, but accuracy varies with the language and the audio quality. Quiet rooms and clear speech help in every language. Because the assistant tells you what it heard first, you can always spot a mistake.

What are the limits?

  • 5 minutes per note

    A voice note longer than 5 minutes is not transcribed. The assistant tells you, and you can type the command or record a shorter note.

  • 60 notes per day

    Up to 60 voice notes are transcribed per day for a workspace. After that, type your commands until the next day.

  • Your own notes only

    Only voice notes you send to yourself are run. Forwarded voice notes, and voice notes from your customers or contacts, are not run as commands.

  • Same safeguards as typed commands

    Paced sending, daily caps, quiet hours and the pause when you reply yourself all apply to commands you speak.

What happens to the audio?

Voice notes sent to your “Message yourself” chat are sent to a third-party speech-to-text provider (Deepgram) to be transcribed. After transcription, whatsappx.si drops its reference to the audio and keeps only the transcript, so the audio itself is not stored.

Only your own voice notes to yourself are transcribed. Messages and voice notes from other people are not transcribed by the assistant.

Frequently asked questions

Is voice-note command available now?
Yes. Voice-note commands are live for the AI Assistant on your own WhatsApp. Link your number by QR code, switch the assistant on and send yourself a voice note of up to 5 minutes. Spoken commands run the same way as typed ones.
Does it transcribe my customers’ voice notes?
No. Only voice notes you send to yourself, in your own “Message yourself” chat, are transcribed and run. Voice notes from customers or contacts, and forwarded voice notes, are not run as commands. Your shared inbox is unchanged.
Does it understand Hindi and Hinglish?
Yes. Hindi and speech that mixes Hindi and English are among the languages it handles best. It detects the language automatically, so you just speak. It also works in other languages, with accuracy that depends on the language and the audio quality.
Is there a limit on voice notes?
Yes. A voice note can be up to 5 minutes, and up to 60 voice notes are transcribed per day for a workspace. If a note is too long or the daily limit is reached, the assistant tells you and you can type the command instead.
What if it mishears me?
It replies with what it heard (🎤 "...") together with what it did, so you can spot a mistake. If it misheard, pause the assistant or send a correction by voice or text. Clear speech and a quiet place improve accuracy, whatever the language.
Do you keep my voice recordings?
No. After transcription, whatsappx.si drops its reference to the audio and keeps only the transcript. The audio is sent to a third-party speech-to-text provider, Deepgram, to be transcribed.

Bring your team into one WhatsApp inbox

Create a workspace, connect your WhatsApp by QR code, and invite your teammates.