Hyprcore

Hyprcore is a native macOS workspace designed to centralize knowledge management through artificial intelligence. This tool integrates three critical functions into a single interface: an LLM-powered smart wiki, an advanced meeting recorder, and a global dictation system. Its primary goal is to empower professionals and teams to capture every spoken or written word, automatically organizing it into a page tree that is accessible and searchable using natural language.

The global dictation feature allows you to write at the speed of thought in any application on your computer. By activating a keyboard shortcut, users can speak and watch their words instantly transform into text wherever the cursor is located, without needing to switch windows. From a technical standpoint, Hyprcore stands out by utilizing up to seven GPU-accelerated local speech engines, such as Whisper, NVIDIA Canary, Parakeet, or Moonshine. This ensures that processing remains private and extremely fast, even offering live translation as you speak.

Regarding session management, the app uses a floating recorder that captures both system and microphone audio on platforms like Zoom, Slack, or Google Meet. The system doesn't just transcribe; it applies speaker diarization to identify who is speaking at any given time. After the session, the AI automatically generates executive summaries, to-do lists, and key decisions, transforming every call into a structured page within the wiki.

The Hyprcore wiki acts as a digital brain where all content is indexable. It utilizes a page structure with icons and cover images, allowing information to be organized visually and hierarchically. One of its most powerful capabilities is the AI Knowledge Base, which enables semantic queries across your entire history. Users can ask specific questions about past meetings, and the tool will respond based on stored transcripts and notes, citing the original sources.

Key technical and functional features include:

• GPU-accelerated dictation with multiple local engines to ensure data privacy.
• Meeting recording with automatic generation of summaries and editable action items.
• ⌘K Universal Search to jump instantly between pages, notes, and meeting transcripts.
• MCP (Model Context Protocol) integration, allowing accumulated knowledge to connect with external AI agents like Claude or Cursor.
• Flexible post-processing, letting you choose between local models like Ollama or cloud services like GPT-4, Claude, or Gemini to refine text and summaries.

Hyprcore is optimized for Apple Silicon and is positioned as a comprehensive solution for users who need to turn conversations and spontaneous thoughts into structured, actionable documentation, while always maintaining control over their privacy by prioritizing local on-device processing.