VoiceHop, a browser extension that translates spoken audio and video in real time, shipped version 2.0.0 on July 28, 2026. The tool sits on top of any website and produces sub-second, speech-to-speech translation of YouTube videos, Netflix titles, Zoom calls, and Google Meet meetings, with voice-preserving output in eight languages and text translation across more than 20. It is a small, indie launch, 701 active users and a 3.8-star rating on the Chrome Web Store, but the capability it packages is one that used to require a full production pipeline or an enterprise subscription. This is a look at what VoiceHop does, how real-time translation differs from the async dubbing tools most creators already use, and where it fits in a multilingual workflow.
What VoiceHop Actually Does
VoiceHop is a client-side Chrome extension. Once installed, it listens to the audio stream of whatever tab you are in and generates a translated voice track on top of it, aiming for latency under one second so the translation keeps pace with the conversation or video. The company describes the output as natural neural voice that preserves tone and style, and it offers voice-preserving translation in eight languages with text-level support for more than 20.
The important detail is where it runs. Because it is a browser extension rather than a platform, it works on any site with audio or video, from a Korean livestream to a French webinar to a client call. There is no upload step, no render queue, and no file to download. You turn it on and listen. Pricing is usage-based, from $0.13 per minute on the Starter tier down to $0.09 per minute on the Scale tier, with a 20 percent discount on annual billing.
How Real-Time Translation Compares to Async Dubbing
Most creators reaching for "AI translation" today actually mean async dubbing: you upload a finished video, the tool renders a translated version, and you download it minutes later. That is what HeyGen and ElevenLabs Dubbing do, and they do it with far broader language coverage and higher polish than any real-time tool can. VoiceHop is solving a different problem: translating live or unowned content on the fly, where you cannot upload a file because the audio is streaming past you in real time.
The other real-time player is Google, whose speech translation in Google Meet reached general availability for businesses on January 27, 2026. It is genuinely real-time and preserves the speaker's cadence, but it only works inside Meet and requires an eligible Google AI subscription. Here is how the four approaches line up.
| Tool | Type | Real-time speech-to-speech | Voice-preserving | Languages | Where it works | Pricing |
|---|---|---|---|---|---|---|
| VoiceHop 2.0 | Browser extension | Yes, sub-second | Yes (8 languages) | 20+ text, 8 voice | Any site: YouTube, Netflix, Zoom, Meet | $0.09 to $0.13 per minute |
| Google Meet Live Translate | Built-in meeting feature | Yes | Yes, mimics tone | English plus 5 at GA, expanding | Google Meet only | Google AI Pro or Ultra |
| HeyGen | Web platform | No, async render | Yes, plus lip-sync | 175+ | Uploaded video files | From $24 per month |
| ElevenLabs Dubbing | Web platform | No, file-based | Yes | 90+ | Uploaded audio or video | From $5 per month |
The takeaway is that these tools are not really competitors. HeyGen and ElevenLabs win decisively for producing a polished, translated version of your own finished video, with lip-sync and dozens more languages. VoiceHop and Google Meet win when the content is live and you do not own the source. VoiceHop's edge over Meet is that it is not locked to one platform or one company's subscription.
How to Translate a Video or Call in Real Time
The workflow is deliberately thin, which is the point. To translate a live stream or call:
- Install the VoiceHop extension from the Chrome Web Store and sign in to create an account.
- Open the tab with the audio you want translated, a YouTube video, a Zoom call in the browser, or a Google Meet room.
- Click the VoiceHop icon and pick your source language and target language. Choose voice-preserving mode if your pair is one of the eight supported.
- Start playback or join the call. VoiceHop generates the translated voice track live, layered over the original.
- Adjust volume balance between the original and the translation so you can still hear speaker cues and tone.
Because processing is client-side and billed per minute, there is no batch step to wait on and no stored file to manage. For a creator, that means you can sit in on a foreign-language product launch or a source interview and follow it live, rather than waiting for a captioned recording the next day.
What This Enables for Creators
Real-time translation unlocks a set of tasks that async dubbing simply cannot touch. You can research a topic by watching non-English tutorials, streams, and press events as they happen. You can run a client or collaborator call across a language barrier without booking a human interpreter. You can monitor a live launch from a studio in another country and react to it the same hour, which matters if your content depends on being early.
It also lowers the localization ceiling for smaller creators. Voice-preserving output means the translated track can carry your own vocal character rather than a flat synthetic read, the same shift that free voice-cloning tools like Fish Audio S2.1 Pro brought to text-to-speech. For a fuller comparison of the production-grade dubbing options, HeyGen's own breakdown of dubbing tools is a useful starting point, and Google's writeup on how it built Meet translation explains why real-time speech-to-speech is technically harder than file-based dubbing.
The Catches: Cost, Quality, and Privacy
A 3.8-star average across a small review base tells you this is early software. Real-time translation trades accuracy for speed: a sub-second pipeline cannot re-read context the way an async renderer can, so expect the occasional mistranslation or clipped phrase, especially on technical jargon or crosstalk. The per-minute pricing also adds up fast for heavy users. A creator watching two hours of foreign-language streams a day at $0.09 per minute is looking at roughly $324 a month, which is well past a flat platform subscription like HeyGen's plans if you only need finished-video translation. VoiceHop earns its cost only when the content is genuinely live or unowned. Privacy is a plus here: processing is client-side, so audio is not necessarily shipped to a server, though you should confirm the current data policy before running it on confidential calls.
Frequently Asked Questions
Is VoiceHop free?
No. It uses usage-based pricing, from $0.13 per minute on the Starter tier (150 minutes) down to $0.09 per minute on the Scale tier (1,100 minutes), with a 20 percent discount on annual billing. There is no fully free tier for unlimited translation.
What languages does VoiceHop support?
It supports more than 20 languages for text-level translation and voice-preserving translation in eight of them, meaning the translated audio keeps a natural, tone-matched voice rather than a generic synthetic one.
Does it work on Zoom and Google Meet?
Yes, on the browser versions. Because VoiceHop is a Chrome extension that reads any tab's audio, it works on browser-based Zoom and Google Meet calls, as well as YouTube, Netflix, and general web video.
How is this different from HeyGen or ElevenLabs?
HeyGen and ElevenLabs are async dubbing platforms: you upload a finished video or audio file and download a translated version, with broader language coverage and lip-sync. VoiceHop translates live, streaming audio in real time with no upload, which is the opposite use case.
Is real-time translation accurate enough for professional use?
For following along, research, and informal calls, yes. For anything where a mistranslation carries legal or contractual weight, real-time AI translation is not a replacement for a certified human interpreter, since the sub-second pipeline sacrifices some accuracy for speed.
Does VoiceHop send my audio to a server?
The company describes the extension as privacy-first with client-side processing. That reduces exposure compared to upload-based tools, but you should verify the current data handling policy before using it on confidential or regulated conversations.
