OpenTypeless

OpenTypeless provides desktop‑wide voice input that converts spoken language into polished text and inserts it directly into any active application. Users invoke a global hotkey, speak naturally, and the software streams transcription while applying AI‑driven polishing, optionally translating into another language or using a custom dictionary for domain‑specific terminology. It supports multiple speech‑to‑text providers (including Deepgram, AssemblyAI, Whisper, Groq, GLM‑ASR, and SiliconFlow) and several large language models for polishing (such as OpenAI, DeepSeek, Claude, Gemini, and Ollama), allowing per‑app formatting and context‑aware output.

The tool is cross‑platform, running on Windows, macOS, and Linux, and can be configured to simulate keyboard input or place results on the clipboard. Features include a floating capsule widget, per‑app detection, local history with full‑text search, dark/light/system themes, and auto‑start on login. It is released under the MIT license, is self‑hostable, and offers a free‑tier configuration that combines Groq’s Whisper‑large‑v3‑turbo for transcription with Google Gemini‑2.5‑flash for polishing.