Windows 11 ships with a built-in LiveCaptions feature that provides on-device speech recognition for accessibility. LiveCaptions Translator builds on this foundation to provide real-time translation of speech captured by LiveCaptions, routing it through various translation backends including large language models (LLMs) and traditional translation services. This approach sidesteps the need for expensive cloud speech-to-text (STT) services or specialized hardware, turning any Windows 11 PC into a multilingual translator.
What LiveCaptions Translator does and how it works
LiveCaptions Translator is a Windows desktop application written in C# that acts as a bridge between Windows 11’s native LiveCaptions speech recognition and multiple translation APIs. It captures the on-device generated captions from LiveCaptions and forwards them to configurable translation backends. These include self-hosted LLMs such as Ollama, OpenAI-compatible APIs, OpenRouter, as well as traditional services like Google Translate and DeepL.
The app features a modern Fluent UI design supporting both light and dark themes, and a borderless overlay window for displaying translated subtitles on top of games or videos, providing an immersive experience without distraction. It also maintains log cards for context tracking and stores full translation history with options to export it as CSV files.
Under the hood, it requires Windows 11 version 22H2 or later, and .NET runtime 8.0 or higher. The critical architectural advantage is leveraging the existing Windows LiveCaptions engine for speech recognition, which operates on-device and is resource-efficient. This avoids the need for cloud STT calls, reducing latency, privacy concerns, and cloud costs.
Technical strengths and design tradeoffs
What distinguishes LiveCaptions Translator is its clever reuse of the Windows LiveCaptions accessibility feature. Instead of building its own speech recognition pipeline or relying exclusively on cloud STT services, it taps into the captions generated locally by Windows. This means the heavy lifting of speech-to-text conversion is offloaded to the OS, which is optimized for low latency and privacy.
The app’s architecture is modular regarding translation backends, supporting a wide range of engines. Notably, it integrates with self-hosted LLMs like Ollama, which can run locally, allowing users to avoid cloud dependencies entirely if desired. This flexibility is valuable for users with privacy concerns or limited internet connectivity.
The Fluent UI implementation provides a polished user experience with support for dark mode and overlay positioning, which matters for usability during gaming or video consumption. The overlay window is borderless and can be positioned over any application, ensuring subtitles remain visible without interfering.
Tradeoffs include the strict requirement for Windows 11 22H2 or later due to reliance on the LiveCaptions feature. Users on older Windows versions or other OSes cannot benefit. Also, while the app supports multiple translation backends, the quality and latency depend heavily on the chosen service. Cloud APIs introduce network latency and potential costs, while local LLMs require sufficient hardware.
From a code quality perspective, the project is actively maintained with a clear separation of concerns: capturing LiveCaptions output, managing translation API clients, UI rendering, and history management. The codebase is written in modern C# with .NET 8.0, making use of the latest language features and libraries.
Getting started with LiveCaptions Translator
Prerequisites
| Requirement | Details |
|---|---|
| Windows 11 22H2 or later | With LiveCaptions support. |
| .NET runtime 8.0 or higher | Recommended. Not tested on earlier versions. |
This tool leverages Windows LiveCaptions introduced in Windows 11 22H2.
Getting started
⚠️ IMPORTANT: You must complete the following steps before running LiveCaptions Translator for the first time.
For detailed information, see Microsoft’s guide on Using live captions.
Step 1: Verify Windows LiveCaptions availability
Confirm LiveCaptions is available on your system using any of these methods:
- Toggle Live captions in the quick settings
- Press Win + Ctrl + L
- Access via Quick settings > Accessibility > Live captions
- Open Start > All apps > Accessibility > Live captions
- Navigate to Settings > Accessibility > Captions and enable Live captions
Step 2: Configure LiveCaptions
When you first start, Windows LiveCaptions will ask for your consent to process voice data and prompt you to download language files used by on-device speech recognition.
After launching Windows LiveCaptions, click the ⚙️ gear icon to open settings, then select Position > Overlaid on screen.
⚠️ VERY IMPORTANT! Otherwise, a display bug will occur after hiding Windows LiveCaptions.
After configuration, close Windows LiveCaptions.
Verdict
LiveCaptions Translator is highly relevant for Windows 11 users who want real-time speech translation without relying on cloud speech-to-text services or specialized hardware. Its use of the built-in LiveCaptions engine is an efficient and privacy-conscious design choice.
The project shines for those comfortable with configuring translation backends, especially if you want to experiment with self-hosted LLMs like Ollama or use popular APIs like OpenAI, Google Translate, or DeepL. The overlay feature and Fluent UI polish make it usable during gaming or video watching, which is a practical use case.
Limitations include the mandatory Windows 11 22H2+ environment and .NET 8.0 runtime, which restricts the user base. Also, the translation quality and latency depend heavily on the translation backend chosen. If you rely on cloud APIs, this might introduce lag or costs.
Overall, it’s a solid, well-architected tool for real-time multilingual subtitles on Windows, worth exploring if your workflow involves live speech translation or accessibility enhancements.
Related Articles
- LiteRT-LM: Google’s C++ library for efficient edge language model inference — LiteRT-LM is a Google AI Edge C++ library for performant language model inference on edge devices with multi-language AP
- Ollama: a unified CLI and API platform for local large language models — Ollama simplifies running and managing open-source large language models locally with a unified CLI and REST API, suppor
- Langfuse: Simplifying LLM observability with decorator-based tracing — Langfuse provides end-to-end observability for LLM applications with automatic tracing via an @observe() decorator, enab
- Jan: a local-first desktop app for large language models with Tauri and Rust — Jan is an open-source desktop app that runs large language models locally using Tauri, Node.js, and Rust. It offers priv
- Voice-Pro: chaining Whisper, translation, and voice cloning in a portable Gradio app — Voice-Pro bundles Whisper variants, translation, and zero-shot voice cloning into a single Python Gradio app, balancing
→ GitHub Repo: SakiRinn/LiveCaptions-Translator ⭐ 2,991 · C#