You stepped into a one-hour meeting. You came back to 62 messages in the work group and six voice notes, one of them five minutes long. Somewhere in there is the only thing that matters: who is sending the proposal, and by when.
Listening to all of it costs the rest of your morning. Not listening costs more later.
Why the group is always full again
Voice became the default way to talk. WhatsApp said so itself in March 2022: people send 7 billion voice messages a day, on average. And voice is not a niche habit — in Brazil, one of the app's largest markets, an Opinion Box survey from June 2025 found that 97% of users open WhatsApp every day and 80% send voice messages.
Reading, though, is faster than speaking. Marc Brysbaert's meta-analysis in the Journal of Memory and Language (2019) put silent reading of non-fiction at 238 words per minute and reading aloud at 183 words per minute. On one message the gap is nothing. On a group with six voice notes it is half an hour.
And there is a third cost nobody counts: the scrolling you do to find where the useful sentence was.
What you can do today, and where each option stops
| Option | Where it stops |
|---|---|
| Speed the audio up WhatsApp plays voice messages at 1.5× and 2×, announced in March 2022. | It is still listening, one message at a time. In a room with other people you still need headphones. |
| WhatsApp's own transcripts Announced in November 2024, they turn a voice message into text. | You have to switch them on in Settings, they work one message at a time (long press, then "transcribe"), and they are generated on your device, with a limited set of languages. |
| Scroll and read everything Nothing to install. | It handles the text and ignores the audio. And a busy group has more messages than patience. |
| Summarize the whole chat What ZapVox does on WhatsApp Web: it takes the text plus the voice messages already transcribed and returns a summary that separates you from the other side. | It needs your computer with WhatsApp Web open, and it covers the last 24 hours. A voice message that has not been transcribed yet is left out — and counted on screen, never dropped silently. |
Three ways to summarize a chat on WhatsApp Web
All of them live on your computer, inside WhatsApp Web, with ZapVox installed. None of them joins the group as a participant: nobody but you knows a summary exists.
Open the chat. At the top, next to the contact's or group's name, click the ZapVox gear and pick 🎧 Audios. It gathers the voice messages already transcribed in that chat — including the ones far up the screen — and returns a summary of those alone.
In the same gear, pick 💬 Whole chat. Now text messages and voice messages from the last 24 hours go in, separated by who said what: your side on one block, the other side on another. In a group, each participant shows up with their name and number. At the end, when there are any, come the to-dos and agreements — "she sends the proposal by Friday".
Click the ZapVox icon in the Chrome toolbar, open Chat tools and use Summarize all audios. Same result as the 🎧 shortcut, by way of the panel.
The summary shows up on a card in the corner of the screen, without writing anything into the conversation. One click on the text copies all of it — to paste into the team email, into your note for that contact, or into your own system.
What a real group summary looks like
An example of the shape, with names changed. The internal group of a small clinic, on an ordinary morning:
🙋 You: confirmed the Thursday schedule change and asked for the quote on the new equipment by Friday.
👤 Ana (+55 49 99999-0000): said the 2pm patient moved to Monday and sent the insurer's rate in a voice message.
👤 Bruno (+55 11 98888-0000): asked for the supplier's contact and will place the order on Thursday.
✅ To-dos and agreements: • Friday — send the equipment quote (You) • Thursday — place the order with the supplier (Bruno)
Note what it does not do: it invents no names, it does not turn a mention into a code, and it skips the "good morning". Dates, times and amounts come out exactly as they were said.
Before summarizing: have the audio ready
A summary reads what has already become text. So the real work happens earlier, and on its own: in the same per-chat gear there is an option to transcribe automatically the voice messages that arrive there. Turned on in the group that matters, every voice note shows its text in the box right below it, with no click from you.
Turn it on in the work group and leave it off in the family group, and you reach the end of the day with the right audio already in text — and your quota spent where it pays. That is the difference between summarizing a chat full of text and summarizing a chat full of "audio not transcribed yet".
The opposite works too: if you would rather transcribe only when you need it, leave everything off and click ✍️ on the message that matters. The summary behaves the same; it just has less material to look at.
Big group: who said it is spelled out
In a ten-person group, "somebody asked for the quote" is useless. That is why every participant enters the summary with their saved name and number beside it — "Ana (+55 49 99999-0000)" — and mentions appear with the name of the person tagged, not the internal code WhatsApp uses underneath. When the mention is you, the summary writes "@You".
It sounds like a detail. In practice it is what separates a summary you forward to the team from a summary you have to check message by message before trusting it.
Where this stops
- It runs on a computer. ZapVox is a Chrome and Edge extension and works inside WhatsApp Web. Not on the phone.
- The window is 24 hours. On purpose: months of conversation would produce a vague summary. When WhatsApp hands over only part of the period, the card says from what time it managed to read.
- Untranscribed voice messages stay out. They are counted on the card — "2 audios not transcribed yet were left out" — so you can decide whether to transcribe first and run it again.
- There are quotas. The Free plan does up to 10 cloud transcriptions a day and 15 minutes of audio; Starter, 60 operations per feature type and 30 minutes; Pro, 150 per type and 150 minutes. Summarizing a whole chat is a Starter, Pro and 5-day trial feature.
- The text goes to the AI to become a summary, and ZapVox does not keep that text. Before your first chat summary, the extension says so on screen and waits for your go-ahead.
When this changes the day
The front desk coming back from lunch, needing to know which patients rescheduled. The lawyer opening a case group with 40 messages who only wants what changed about the deadline. The agency with one group per client, and a client who explains everything by voice. The new parent in the school group, where the useful line hides between thirty stickers.
In all of them the question is the same: what do I need to know, and what was agreed? That is the question a summary answers.
Try it free for 5 days, no card. Transcribe, translate and summarize the whole chat on WhatsApp Web.