Skip to content

lcode

A coding agent that runs entirely on your own machine. Open-weight models · up to 256K tokens of context · your NVIDIA GPU or Apple Silicon Mac · your code stays local.

Get started View on GitHub

curl -fsSL https://nasser1941.github.io/lcode/install.sh | bash

lcode fixing a failing test

What it does

Point lcode at a repository and talk to it in your terminal. It explores the code, answers questions with file:line references, writes and edits files, runs your scripts and tests, and keeps going until the task is done, asking before it changes anything.

  • Private by design

    The model runs locally through Ollama, lcode never uploads your files and sends no telemetry. Web search is there when the model needs current information, and can be turned off.

  • A real agent, not autocomplete

    Reads, searches, edits and runs commands in a loop, verifies its work, and plans multi-step tasks with a todo list.

  • Long context

    Up to 256K tokens with the default model (1M with Nemotron), so it can hold a large part of your repository in view. You choose the window size.

  • Hardware-aware

    Detects your GPU or Mac and picks the best model and the largest context that fits: lcode setup does it in one step.

  • Safe by default

    Every edit is shown as a diff and every non-read-only command needs your approval, unless you choose otherwise. /undo takes back a request's file changes.

  • MCP servers

    Jira, GitHub, AWS, Google Drive, Grafana and more, one command each: lcode mcp add atlassian. Browser sign-in is built in.

  • Bring your own model

    Eight curated open-weight models out of the box, and any other Ollama model with tool calling works too.

Runs on

Platform Recommended hardware Default choice
Ubuntu / Linux with NVIDIA GPU 12 GB+ VRAM and 32 GB RAM Qwen3.6 35B-A3B at 256K context (50–55 tok/s on 12 GB)
macOS on Apple Silicon (M1–M4) 48 GB+ unified memory Qwen3.6 35B-A3B at 256K context (45 tok/s on an M4 Pro)
Smaller machines (16–24 GB Mac, 8 GB GPU) — Qwen3.5 9B / 4B with a smaller context

See Models & context windows for the full catalog and how lcode sizes models to your hardware.

Next steps