gladiaflow
An open-source desktop application that transcribes real-time speech via Gladia API and pastes text into active applications.
About gladiaflow
GladiaFlow provides instant, cross-platform voice dictation, allowing users to stream speech-to-text directly into any desktop application with a single global hotkey.
For the Non-Technical Reader
Think of GladiaFlow as a universal speech-to-text bridge attached directly to your text cursor. Instead of switching between transcription apps or relying on OS dictation tools, you simply hold a hotkey (like Fn/Globe on macOS or Ctrl + Space on Windows) and speak. As soon as you finish, polished, properly punctuated text appears directly inside your active window, whether that is Slack, VS Code, Google Docs, or an email client.
For the Technical Reader
GladiaFlow is an open-source cross-platform desktop application built on top of Tauri 2, Rust, and a Node.js/Vite frontend.
Audio Streaming Pipeline: Captures raw PCM audio chunks from the device microphone and streams them live via a low-latency WebSocket session to the Gladia Live Transcription API.
Transcript Processing: Receives partial and final transcription events, passing final utterances through a cleanup pipeline that manages capitalization, automatic endpointing, punctuation, and deduplication.
Native Desktop Integration: Utilizes Rust native capabilities for global hotkey handling (push-to-talk or toggle mode), local history management, system tray integration, and cross-application text injection/pasting.
Customization & Multi-Language Support: Features built-in support for custom vocabulary, domain-specific pronunciations, language hints, and real-time code switching.
Why It Matters
By pairing a lightweight native runner (Tauri/Rust) with specialized cloud transcription APIs, GladiaFlow provides a performant alternative to resource-heavy Electron apps or closed proprietary dictation software. It brings enterprise-grade transcription, including custom jargon and multi-language support, directly to the user desktop while retaining local control over application history, privacy settings, and workflow statistics.
The "Voice AI Space Lab" Idea
The Hands-Free Terminal Co-Pilot: Pair GladiaFlow's push-to-talk workflow with a local LLM shell interpreter. Developers can hold a hotkey, describe complex shell operations or multi-step Git workflows in natural language, and watch precise CLI commands auto-fill and execute in their terminal window in real time.
Explore the repository on the GladiaFlow GitHub Page.