ZapVox VS JavaScript Tutorial

ZapVox's answer

• A Chrome Manifest V3 extension written in TypeScript, which also runs on Edge and other Chromium browsers
• Supabase (Postgres plus Edge Functions) as the backend and orchestration layer
• Speech-to-text and language models from Groq (Whisper), Google Gemini and OpenAI, with the option to bring your own key
• Audio decryption done locally in the browser with HKDF-SHA256 and AES-256-CBC, the same scheme WhatsApp uses internally
• Vercel for the website and Stripe for billing

ZapVox's answer

Most transcription tools ask you to leave the conversation: export the audio, upload it somewhere, wait, come back. ZapVox works where the message already is — a Transcribe button sits on every voice message inside WhatsApp Web, and the text appears right under the audio bubble.

It is also more than a transcript. The same click can translate the message into 99+ languages, summarize a long recording, or read the tone behind it — and it works in groups, not just one-to-one chats.

The audio is decrypted inside your own browser before anything leaves it, and it is never stored on our servers.

ZapVox's answer

It stays inside WhatsApp Web. There is nothing to export and no second tab to open - the button sits on the message itself, and the text lands under the audio bubble.

It also does the whole job from that one click: the transcript, translation into 99+ languages, a summary when the recording runs long, and a read of the tone behind the message. Most alternatives are built for meetings and calls, not for the chat where the audio actually arrives.

The audio is decrypted inside your own browser and is never stored on our servers, which matters when the messages come from clients, patients or your own team.

And the price is friendly: a free tier with no credit card, and Pro from R$ 29.90/month (about US$ 5.50) with a 7-day trial.

ZapVox's answer

People whose work arrives as voice messages and who cannot always stop to listen to them.

In Brazil, where WhatsApp is the default channel for business, that means a lot of sales, support, real estate, legal and healthcare work, plus freelancers and small business owners who get dozens of audios a day and need the content, not the recording.

The second group is anyone working across languages, where translating the message matters as much as reading it, and anyone who is simply on the receiving end of long voice messages from family and friends.

ZapVox's answer

ZapVox started from a very Brazilian problem. WhatsApp is where work happens here, and a large share of it arrives as voice messages - often several minutes long, often at the worst possible moment: in a meeting, in a noisy place, in the middle of something else.

The tools that existed all asked for the same detour: export the audio, upload it to another service, wait, come back with the text. That is fine once. It is unworkable when it happens twenty times a day.

So the goal was narrow from the start: put the transcript where the conversation already is, one click away, and then make it useful - translated, summarized, with the tone made explicit - without ever storing anyone's audio.