Real-time speech translation,
right on your desktop.
Capture any audio playing on your PC or microphone, translate it instantly with NVIDIA Riva, Whisper, or Llama 3.1, and watch it appear as a transparent always-on-top overlay. No extra hardware. No browser tabs.
Features
Everything you need, nothing you don't
Built for live captions that follow what you're actually doing — gaming, meetings, foreign-language content, accessibility.
Universal audio capture
Translates any audio playing on your PC — videos, calls, streams, meetings — or your microphone. One-toggle switch with independent volume controls.
Multi-engine ASR
Pick NVIDIA Riva for highest accuracy, Whisper for offline privacy, or Google for zero-setup speed. Switch any time without restarting.
Multi-engine translation
Llama 3.1 8B for context-aware AI translation, Google Translate for breadth, or Riva NMT for neural quality. Bring your own NVIDIA key for unlimited.
Transparent overlay
Always-on-top, draggable, resizable. Adjustable opacity, font size, bold toggle, and a Mini Mode that collapses to a single caption line.
Searchable history
Every caption is saved to a glassy dark History Panel — search, jump back, and copy any line. Retention scales with your tier.
Account & sync
Sign in with Google, email, or as Guest. Settings sync across devices. Live usage dashboard, real-time quota tracking, Razorpay-powered billing.
Engines
Pick the engine that fits the job
Speech recognition and translation are decoupled — mix any pair. Bring your own NVIDIA key to bypass quotas on Riva and Llama engines.
Speech recognition
3 enginesTranslation
5 enginesHow it works
From install to live captions in minutes
No configuration wizards. No browser extensions. Install, sign in, pick your engines, done.
Install
Download OmniBridge_Setup.exe from the Releases page. The installer bundles the Python server and Flutter UI — no separate setup, no Docker.
Sign in
Sign in with Google or email — or continue as Guest. Settings sync to the cloud the moment you log in.
Pick engines & language
In Settings → choose a speech recognition engine, a translation engine, and your target language. Defaults work out of the box.
Hit play
Captions appear live as audio plays on your PC. Drag the overlay anywhere, resize it, collapse it to Mini Mode for a single line.
Pricing
Free forever, scale when you need it
Token-based quotas across all engines. Bring your own NVIDIA API key to bypass quotas on NVIDIA-backed engines on any tier.
Free
Get started with live captions. Daily quota resets every 24h.
- 5,000 tokens / day · 50,000 / month
- Google Translate & MyMemory
- Google Online ASR
- 15-minute session limit
Pro
Everyday users. All translation engines, Whisper Tiny–Small, microphone audio.
- 25,000 tokens / day · 250,000 / month
- All translation engines
- Whisper Tiny / Base / Small
- Microphone capture
- 7-day caption history
- 2-hour sessions, 2 concurrent
Enterprise
Power users & teams. Highest accuracy ASR, longest sessions, generous concurrency.
- 75,000 tokens / day · 750,000 / month
- Whisper Medium + NVIDIA Riva
- 30-day caption history
- 8-hour sessions, 5 concurrent
- Priority support
Download
Built for Windows desktop
One installer, no separate runtime to manage. Updates land via the in-app updater — no store, no sideloading.
Single Inno Setup installer. Bundles the Python server and Flutter UI — survives upgrades, keeps your Whisper models, and signs you back in automatically.
Available nowAny modern NVIDIA card unlocks low-latency Riva ASR / NMT engines. CPU-only systems still work — engines fall back to cloud or Whisper.
Available nowCross-platform builds aren't on the v2 roadmap. If you'd like to see them, open a GitHub issue with your use case.
Tracking interest