Transcribe a WhatsApp voice message without scanning the chat

Start from the message you select

Open the message menu on a single voice note and choose transcription. The extension reads that message's audio and sends it for recognition.

It does not scan other messages, other chats, or your contact list. If you need three voice notes transcribed, select three times.

• One selected voice message per request.
• No batch scanning of the chat.
• No access to messages you did not select.
• The result is attached to that message.

Formats, timing, and language

Supported audio formats are OGG/Opus, MP3, WAV, and M4A. Other formats are rejected rather than converted silently, so you know why a file did not work.

Voice recognition runs asynchronously: the task is submitted, then the result arrives when processing finishes. You can keep working in the chat while it runs.

Language detection is automatic by default. When the audio is unclear or mixes two languages, the result may include the detected language so you can judge it.

Transcription can be paired with translation: recognize first, then translate the recognized text. The two steps answer different questions.

Transcription is not translation

A transcript keeps the speaker's original words. A translation changes them into your language. Neither one is a summary, and neither one adds facts that were not said.

If you need both, run recognition and then translation. If you only need to know what was said, recognition is enough and costs less.

Keep the two results separate in your notes. Mixing a transcript with a translation makes it impossible to quote the speaker accurately later.

Accuracy depends on the audio

Recognition can be wrong. Background noise, overlapping speakers, strong accents, poor microphone quality, and technical vocabulary all reduce accuracy. Numbers, model codes, and names are the most fragile.

Treat the transcript as a draft. Before you rely on a quantity, a price, or a date, listen to that part again or ask the sender to confirm in writing.

Some audio cannot be processed at all: a message that will not play, a file in an unsupported format, or a recording with no usable speech. The tool reports the failure instead of inventing text.

Privacy and storage boundaries

The extension processes only the voice message you select. It does not read your chat history, your contacts, or your other media.

Derived results are stored locally in the extension by message ID and source fingerprint. They expire after 90 days, are capped at 500 entries, and can be cleared in settings. Local storage is not a vault and is not claimed to be encrypted.

Voice files and transcripts may contain personal or commercial information. Check your own retention policy before sharing a transcript outside your team.