Live dubbing
Hear any video spoken in your language while it plays. Natural AI voices keep up with the speaker.
Free & open sourcev1.1.17
Hear every voice in your language.
Live AI dubbing and bilingual subtitles for any video — in your browser, on your desktop or from a file. Understand courses, films, podcasts and streams in the language you think in.
- No ads, no analytics
- Your keys stay with you
- 79 target languages
- Chrome · Edge · Windows · macOS · Linux
Original
Waiting for audio…
Translate into
Tap a language and watch the voice switch.
Made for
- Online courses
- Films & series
- Podcasts
- Lectures
- Livestreams
- Documentaries
- Audio & video files
Features
Everything you need to understand any video
Dub it, caption it, mix it, export it — every piece works on its own.
Dual subtitles
Original and translated captions side by side, on a frosted-glass card you can drag, resize and fade.
Four independent outputs
Original audio, dubbed audio, source subtitles, translated subtitles — switch each one on or off.
A mixer you control
Balance the original against the dub, with smart ducking that lowers the original while the dub speaks.
Synchronized player
Capture runs ahead while an independent player pauses, seeks and goes fullscreen. Export WebM and SRT.
Desktop app & files
Translate any program’s audio live, or drop in an audio or video file and get a synced player, subtitles and a ZIP.
79 languages
Choose your target language from the full Gemini set — the most popular ones come first.
Private by design
No analytics, no ads, no relay servers. Audio goes straight to the AI provider you choose.
Free & open source
MIT-licensed. Bring your own free API keys, read every line, make it yours.
How it works
From install to dubbed in three steps
-
Install & add your key
Add Avorythm to Chrome or Edge, paste a free Gemini API key from Google AI Studio and give consent.
-
Choose language & mode
Pick your target language, then watch right on the page for the lowest delay or in the synchronized player for steady sync.
-
Press “Start translating”
Hear the dub and read both captions in real time. Mix, move, pause and export whenever you like.
Watch
See it in action
A short film, a product demo and a step-by-step tutorial — in your language.
Watch in
Two ways to use it
In your browser or on your desktop
The extension and the desktop app are independent — use either one, or both.
Browser extension
Translate any tab
Add it to Chrome or Edge and dub the video you’re watching. It stands alone — no desktop app, Python, FFmpeg or virtual audio cable.
- “On this page”: the lowest-delay path
- Synchronized player: seek, fullscreen, export WebM + SRT
- Movable frosted-glass subtitles
Desktop app
Translate anything you hear
Dub audio from any desktop program, or drop in an audio or video file and get a synchronized player, subtitles and a ZIP of every output.
- Live translation of any program’s audio
- File Studio for audio and video up to 8 GB
- Windows, macOS and Linux
Tutorial
Up and running in minutes
Follow along here, or watch the tutorial video above.
-
Add Avorythm from the Chrome Web Store and pin it to your toolbar. In Edge, use the same listing — or load the ZIP from the release page.
-
Create a key in Google AI Studio, then open Avorythm → Settings → “Connect to Gemini”, paste it and press Save. Finally tick the consent box so audio can be sent to Google Gemini — nothing is captured before you press Start.
-
Open the popup and pick a target language, then “On this page” (lowest delay) or “Synchronized recorder & player” (steady sync, seeking and export).
-
Open a normal video tab and press Start translating. Hear the dub, read both captions and drag the glass card wherever you like.
-
Switch the four outputs on or off, balance the original against the dub, and set the subtitle size and opacity in Settings.
-
Get the build for your system from Releases. The Windows installer already includes FFmpeg.
-
Open Advanced settings and save a Gemini key (plus a Groq key for file processing). Keys live in your operating system’s keyring.
-
To translate another program live, send its sound to a loopback device — AMM Virtual Audio on Windows, BlackHole on macOS, a monitor source on Linux — and pick it as the app’s audio input.
-
Choose your language and output mix, then press Start translation. Keep the listening output on your real headphones.
-
In the desktop app, open File Studio and drop in an audio or video file (up to 8 GB).
-
Choose the target language and a voice. “Precise” gives tighter sync; “Fast” finishes sooner.
-
Leave Avorythm running while it transcribes, translates, narrates and aligns your file.
-
Watch in the synchronized player, or download original.wav, dubbed.wav, source.srt and translated.srt — or everything as one ZIP.
Languages
79 languages, one click
Pick your target language from the full Gemini set — popular ones first, from Arabic and Chinese to Swahili and Zulu.
- English
- فارسی
- العربية
- 简体中文
- 繁體中文
- Deutsch
- Français
- Italiano
- Español
- Русский
- 日本語
- 한국어
- Türkçe
- Português (Brasil)
- Português europeu
- Nederlands
- Polski
- Українська
- हिन्दी
- اردو
- עברית
- Indonesia
- Melayu
- Tiếng Việt
- ไทย
- Afrikaans
- Akan
- Shqip
- አማርኛ
- Հայերեն
- Azərbaycan
- Euskara
- Беларуская
- বাংলা
- Български
- မြန်မာ
- Català
- Hrvatski
- Čeština
- Dansk
- Eesti
- Filipino
- Suomi
- Galego
- Ქართული
- Ελληνικά
- ગુજરાતી
- Hausa
- Magyar
- Íslenska
- Jawa
- ಕನ್ನಡ
- Қазақ тілі
- ខ្មែរ
- Ikinyarwanda
- ລາວ
- Latviešu
- Lietuvių
- Македонски
- മലയാളം
- मराठी
- Монгол
- नेपाली
- Norsk
- Norsk bokmål
- ਪੰਜਾਬੀ
- Română
- Српски
- سنڌي
- සිංහල
- Slovenčina
- Slovenščina
- Basa Sunda
- Kiswahili
- Svenska
- தமிழ்
- తెలుగు
- O‘zbek
- IsiZulu
The interface is available in English, Persian and Simplified Chinese; dubbing and captions work in all 79 languages.
Privacy
Private by design
Avorythm is built so that nobody — including us — sits between you and your media.
No analytics, ads or telemetry
There is no tracking code in the app, the extension or this website.
Your keys, your device
API keys stay in your browser session by default, or in your operating system’s keyring on desktop.
Straight to the AI provider
Audio goes directly from your device to Google Gemini — and, in precise mode, short windows to Groq Whisper. Never to a server of ours.
Consent first
Capture starts only after you give consent and press Start, and you can stop at any moment.
Open source (MIT)
Every line is public. Read it, audit it, improve it.
Avorythm relies on preview AI services. Accuracy, voices, latency and quotas can change — double-check anything important. Read the privacy policy
Download
Get Avorythm
Free and open source. Choose your platform — you only need your own free Gemini API key.
Chrome & Edge
Install from the Chrome Web Store. Edge users can add it from the same listing.
Windows
64-bit installer with FFmpeg included.
macOS
Native builds for Apple Silicon and Intel.
Linux
x64 build. Choose your PipeWire or PulseAudio monitor source in the app.
Extension ZIP
Manual install for Chrome or Edge with Developer mode.
The Windows installer is unsigned while code-signing approval is pending, so Windows may warn about an unrecognized publisher. Verify your download against SHA256SUMS.txt from the same release.
Version 1.1.17SHA256SUMSAll releasesChangelog
FAQ
Good questions
Is Avorythm really free?
Yes. The extension and the desktop app are free and open source (MIT). They use your own API keys: Google AI Studio offers a free Gemini tier, and Groq’s free tier powers Whisper for files. Provider limits and quotas apply and can change.
Do I need the desktop app to use the extension?
No. The extension is fully standalone — no desktop app, Python, FFmpeg, localhost or virtual audio cable. The desktop app is for translating other programs’ audio and for processing files.
Which sites does it work on?
Any normal web page that plays audio or video in Chrome or Edge 116+. DRM-protected streams and browser-internal pages (such as chrome://) can block capture; use the “On this page” mode when video capture isn’t possible.
How much delay is there?
Live translation has to hear a phrase before it can speak it, so a small delay is unavoidable. “On this page” is the lowest-delay path. The synchronized player records ahead (20 seconds by default) so audio, video and captions stay on one timeline.
Is my audio private?
Capture starts only after you give consent and press Start. Audio goes straight from your browser to Google Gemini — and, in precise mode, short windows to Groq Whisper — never to servers of ours. There are no ads, analytics or telemetry, and keys stay in your browser session unless you choose to remember them.
Which languages are supported?
79 target languages, from English, Persian, Arabic and Chinese to Swahili and Zulu. The interface itself is in English, Persian and Simplified Chinese, with more planned.
How accurate is it?
Avorythm uses preview AI services. Accuracy, voices, latency and quotas can change, and no AI translation is perfect — double-check anything important.
Does it work in Firefox or Safari?
Not yet. Avorythm supports Chrome and Edge (Chromium 116 or newer). Edge users can install it from the Chrome Web Store listing; a dedicated Microsoft Edge Add-ons listing is on the roadmap.
Why is the Windows installer “unsigned”?
A code-signing application with the SignPath Foundation is pending. Until it’s approved, Windows SmartScreen may show an unrecognized-publisher warning. Verify your download against SHA256SUMS.txt from the same release.
How can I help?
Star the repository, share the videos, report bugs, suggest features or send a pull request — everything happens in the open on GitHub.