ashlink95/chilljinn.nvim — explained in plain English
Analysis updated 2026-05-18
Get AI code autocomplete in Neovim without sending code to the cloud.
Have suggestions grounded in your project's actual function and variable signatures.
Fix compiler or debugger errors using the error-aware cj-shell companion tool.
Train the autocomplete to match your personal coding style over time.
| ashlink95/chilljinn.nvim | 6elphegor/warp | adoslabsproject-gif/liara-toolkit | |
|---|---|---|---|
| Stars | 4 | 4 | 4 |
| Language | Rust | Rust | Rust |
| Setup difficulty | hard | easy | hard |
| Complexity | 4/5 | 5/5 | 4/5 |
| Audience | developer | researcher | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Rust, Python for model download, Neovim 0.11+, and compiling a native plugin with optional GPU support.
chilljinn.nvim is a plugin for the Neovim text editor that adds AI code autocomplete, but it runs entirely on your own machine instead of calling out to a cloud service. It uses a local AI model called Qwen2.5-Coder, and there is no telemetry and no internet connection needed once it is set up. What makes it different from a plain autocomplete tool is that it tries to be context aware. As you type, it looks at the real function and variable signatures from your own project, its dependencies, and the programming language's standard library, and feeds that into the model. It also constrains what the model is allowed to generate so it cannot invent a function or variable name that does not actually exist in your code. Under the hood, the plugin is built in Rust. A background process runs the AI model and fills a fast in-memory lookup structure that Neovim's completion system reads from, so the editor itself never freezes while suggestions are generated. Everything the plugin needs to remember between sessions, including the downloaded model and its own settings, is stored in a single local database file. Getting started means cloning the plugin into Neovim's plugin folder, running a small Python setup script to download the AI model, and then compiling the Rust code with an optional flag for GPU support through Vulkan or CUDA, or plain CPU if you skip that. Once installed, you turn it on inside a project with a start command, which loads the model and scans your code, and turn it off again with a stop command. The plugin can also learn your personal coding style: you can tell it to remember snippets of code you like, then ask it to turn those into a short set of style rules that shape future suggestions. There is also a companion tool called cj-shell, a small terminal you run your builds and programs inside. It reads the error messages from compilers and debuggers and feeds them back to the plugin so it can suggest fixes right on the line that caused the problem. No license information is given in the README.
A local, offline AI autocomplete plugin for Neovim that reads your project's real code structure to suggest context-aware, non-hallucinated completions.
Mainly Rust. The stack also includes Rust, Neovim, Qwen2.5-Coder.
No license information is provided in the README.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.