vanilla-closedcode
API Docs Releases Discord GitHub

vanilla-closedcode

a local-LLM-only desktop coding assistant

vanilla-closedcode is a desktop coding assistant that runs entirely against your local LLM. Your code, prompts, and project context never leave your machine — no cloud API keys, no remote inference, no telemetry.

Download API Documentation

100% local inference / Desktop app + CLI / Windows macOS Linux

vanilla-closedcode desktop app: file explorer, code editor, an AI chat driven by a local model (gpt-oss), and a git changes panel

Highlights

Local-LLM-only by design

All inference runs against a model on your own machine or LAN. No cloud providers, no API keys, no telemetry of your code or prompts.

Desktop app + CLI

Use the closedcode command-line agent or the desktop app — both share the same core engine.

Edits, runs, and reviews code

Reads your project, proposes and applies edits across files, and runs tools — with you in the loop for every change.

Modular monorepo

Built as composable packages (core, SDK, app, CLI) so you can embed or extend the agent in your own tooling.

Bring your own model

Point it at any OpenAI-compatible endpoint — Ollama, LM Studio, llama.cpp, or your own server. No vendor lock-in.

English / Japanese docs

API documentation is published in both English and Japanese.

Download

Builds are published on the GitHub Releases page. Grab the installer for your platform from the latest release.

Go to Releases

Getting started

  1. Download and install vanilla-closedcode from the Releases page.
  2. Point it at your local model endpoint in the settings (no cloud keys needed).
  3. Open your project folder, then chat with the agent or run the closedcode CLI from a terminal.
  4. Review and accept the proposed edits — nothing is sent to a remote server.

Documentation

JSDoc-generated API documentation for the source is published in two languages:

See the user manual for usage guides and configuration.

Support

Join the community Discord for help and discussion. Questions, bug reports, and feature requests are also welcome on the GitHub issue tracker, or contact support@informanellica.com.