# Nyx

A local-first AI coding agent living in a VS Code or Cursor sidebar that sends every request to the user's own machines, running against Ollama, LM Studio or a self-hosted OpenAI-compatible endpoint, with no account, cloud connection or per-token cost.

Nyx is a local-first coding agent panel for VS Code and Cursor, built to look and behave like a native agent sidebar: file-edit diff cards the user reviews before they land, checkpoints, approvals, @-mentions, semantic search, MCP tools, a job queue, and project memory. Every request goes to machines the user controls rather than a hosted service, so there is no account, no telemetry and no per-token bill.

It targets recent open-weight coding models such as DeepSeek V4, GLM 5.2 and Qwen Coder, running through a user's own inference stack, for example Ollama, llama.cpp, LM Studio, or a self-hosted OpenAI-compatible server such as vLLM or SGLang on hardware like a DGX Spark, a Mac Studio or a high-VRAM GPU workstation. Nyx's machine manager auto-discovers any of these endpoints and lets a user point the sidebar at whichever machine is running a model.

The only optional network use is web fetching and search, a one-time OCR language download, and the first-time download of a vision or embedding model, all under the user's control.

Website: https://github.com/sthamann/nyx-local-ai