vimtelligence
v0.1.9
Published
An ultra-high-performance modal terminal text editor with local AI acceleration
Readme
Vimtelligence is a next-generation terminal text editor built in pure Rust. It marries the lightning-fast, modal editing paradigms of Vim with a locally hosted, offline-first Artificial Intelligence backend. No cloud APIs. No lag.
By embedding a quantized Gemma 3 (270M parameters) model natively via the Candle ML framework, Vimtelligence understands your natural language instructions and translates them instantly into structural Vim commands right within your editor.
✨ Features
- Blazing Fast: Written in pure Rust using
ratatuifor immediate-mode rendering andropeyfor heavy-duty B-tree text buffer management. - True Modal Editing: Supports classic Vim motions and structural operations like
ciw(change inside word),daw(delete around word), anddd. - EZ Mode: Press
Ctrl+Wto open the AI command overlay. Type natural language (e.g., "delete this word"), and the background AI thread will compute and seamlessly execute the correct macro instantly. - Offline & Private: Inference runs 100% locally on your CPU/GPU using the quantized GGUF model. Your code never leaves your machine.
- Asynchronous Engine: Built on
tokio, the editor's UI remains fully responsive at 60 FPS while the AI generates macros in the background.
🚀 Installation
You can install Vimtelligence globally via NPM. The postinstall script will automatically download the compiled Rust executable for your platform and the quantized AI model.
npm install -g vimtelligenceNote: Currently, pre-compiled binaries are provided for Windows (x86_64-pc-windows-msvc). For other platforms, you will need to build from source.
🛠️ Usage
Simply run Vimtelligence from your terminal, passing the file you wish to edit:
vimtelligence src/main.rsCore Keybindings
i- Enter Insert Mode.Esc- Return to Normal Mode.h,j,k,l- Navigate cursor (Normal Mode).s- Save the current file.q- Quit the editor.
EZ Mode (AI Assistant)
- In Normal Mode, press
Ctrl+Wto open the AI ribbon. - Type a natural language command like "delete the current line".
- Press
Enter. The model will execute the command dynamically and return you to Normal Mode.
🏗️ Building from Source
To build Vimtelligence locally, ensure you have Rust installed.
git clone https://github.com/turtle170/Vimtelligence.git
cd Vimtelligence
cargo run --release -- example.txtTo enable the AI capabilities locally, you will need to download the model into the expected directory:
node install.js📜 License
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
