Some voice messages just can't be played right now. You're on the train without headphones, in a meeting or in a quiet office. Some run for five minutes when all you need is an address, a time or a phone number that you would rather copy than replay three times. And some arrive in a language you only half understand. As text, the message can be skimmed, searched and quoted.
Getting the file out of the messenger takes one step. On Android, install Wovea as an app through Chrome's Install app option; after that it shows up in the share menu of WhatsApp, your gallery and your file manager, so you long-press the voice message, tap Share and pick Wovea. Files shared this way can be up to 95 MB. On iPhone, Safari doesn't let web apps join the share menu: forward or share the voice message and choose Save to Files, then tap Choose a file on the Wovea page. In Telegram on a computer, right-click the message and use Save As to get an .ogg or .oga file.
If the transcription built into your messenger doesn't know your language, or you need a copy of the text in a file, this page helps. The language is detected on its own, and the text can go further: “Translate text” puts it into your language, and “Key points” can be made in another language too. If several voice messages piled up, pick them all at once: each one is transcribed separately and you see the text of each, one after another.
You get the transcript split into phrases, each with its timestamp. Copy it, or download it as TXT, Word (DOCX), SRT or VTT. A one-minute voice message takes a few seconds: in our tests, 65 seconds of audio came back as text in 3 to 4 seconds, plus the upload. Formats include Opus, OGG, OGA, M4A, MP3, WAV, AAC, AMR and FLAC, and the sound from a video works too. A file can be up to 120 minutes and 2 GB, so longer recordings fit as well.
The result depends on the recording. A clear voice close to the phone comes out well. Wind, traffic, music or two people talking at once can cause mistakes, so reread anything important, such as names, amounts and addresses, before you rely on it.
The audio is processed on our server and the file is deleted from there after 24 hours. The text is kept in the library in this browser and goes to your account only if you sign in. To stop bots, one connection can run up to 60 AI tasks an hour.
