johnbean393/KeyType

An open-source Cotypist with macOS system wide AI autocomplete

What it solves

KeyType provides system-wide tab-autocomplete for macOS, allowing users to quickly accept predicted text continuations across any application. It serves as an open-source alternative to closed-source tools like Cotypist.

How it works

The utility monitors the focused text field in any active app. It uses a local LLM (via a llama.cpp wrapper) to predict a short continuation of the text at the cursor. This prediction is presented as ghost text, which the user can accept by pressing the Tab key.

Who it’s for

macOS users who want on-device, AI-powered text completion across their entire operating system.

Highlights

  • On-device processing: Uses local LLMs to ensure privacy and speed.
  • System-wide integration: Works across any macOS application.
  • Caret-aware autocomplete: Monitors focused text fields and snapshots the context to provide relevant predictions.
  • Open-source: Released under the MIT license.

Related

  • Project
    deerwork-ai/deer-workflow

    A dynamic workflow runtime for building observable, reusable agent graphs using a code-first TypeScript approach to control AI agent execution.

  • Project
    MioMioOS/MioIsland

    A macOS native companion for Claude Code that brings agent status monitoring and permission approvals to the Dynamic Island and synchronizes sessions to iPhone.

  • Project
    thiswillbeyourgithub/wdoc

    A powerful RAG system that summarizes and queries diverse document types, including PDFs, audio, video, and Anki cards, providing sourced answers and detailed summaries.

  • Project
    skygazer42/MimirQ

    An observable and replaceable enterprise RAG infrastructure that provides a controllable pipeline for document parsing, chunking, and retrieval with built-in regression testing.