Open-Less/openless

What it solves

OpenLess is a cross-platform voice-input application that eliminates the friction between spoken thoughts and clean, usable text. It moves beyond simple word-for-word dictation by using AI to polish, structure, and format spoken input directly into any active text field on your computer, removing the need for manual editing or coordination between different AI tools.

How it works

Users hold a global hotkey to record audio, which is then processed through a pipeline of Automatic Speech Recognition (ASR) and an AI polish model. The system transcribes the audio and then applies a selected "style pack" (a custom system prompt) to refine the text. The final result is streamed character-by-character into the cursor's current position, with a clipboard fallback if direct insertion is blocked.

Who it’s for

• AI Power Users: Those who frequently write complex prompts for LLMs like ChatGPT, Claude, or Cursor and want to compose them via speech.
• Content Creators: People drafting emails, specifications, or social media posts who prefer speaking over typing.
• Developers: Engineers capturing code comments, commit messages, and PR descriptions quickly.
• Anyone who prefers voice input but finds traditional dictation too messy for professional use.

Highlights

• AI-Prompt Mode: Specifically designed to turn loose, spoken requests into structured, detailed prompts for other AI models.
• Style Pack Marketplace: A community-driven system where users can create, share, and install custom system prompts to control the output tone and format.
• Bring Your Own Model: Supports a wide array of ASR and polish providers (including OpenAI, DeepSeek, Google Gemini, and local models like Qwen3-ASR) via cloud credentials or compatible endpoints.
• Local-First Privacy: Data and dictionaries stay on the machine; the app is fully open-source.
• Streaming Insertion: Reduces perceived latency by flowing text into the cursor in real-time as it is polished.