How do I set up a local-first development environment for offline coding and debugging?

A local-first development environment prioritizes working entirely on your machine so you can code, build, test, and debug without an internet connection. The core idea is to install tools and dependencies locally (or in local containers), cache everything you need while online, and avoid cloud-dependent features. Here is a practical setup that works for most languages and stacks.

1. Install the foundational local tools

Choose a capable offline-friendly editor: VS Code (with extensions installed while online), Neovim, or a JetBrains IDE. Install language runtimes and package managers for your stack (Node.js, Python, Go, Java, Rust, etc.) directly on the host or via version managers such as nvm, pyenv, asdf, or uv. Install Git, a terminal multiplexer (tmux or equivalent), and any debuggers or language servers you rely on. Disable telemetry and automatic updates in the editor so it does not try to reach the network.

2. Use containers for reproducible, isolated environments

Install Docker Desktop (or Docker Engine + Docker Compose on Linux). For VS Code users, install the Dev Containers extension. Create a .devcontainer/devcontainer.json (or a docker-compose.yml) that defines the exact runtime, tools, and services your project needs. Once the container image is built or pulled while you have internet, you can reopen the project in the container and work completely offline. The same approach works with plain Docker Compose for multi-service apps (database, cache, message queue, etc.). Mount your project directory into the container so edits on the host are immediately visible inside the environment.

3. Make dependencies available offline

While online, run a full install of every project’s dependencies (npm install, pip install -r requirements.txt, cargo build, go mod download, etc.). This populates local caches. For heavier use, set up local package mirrors or caches (Verdaccio for npm, devpi for Python, a local Maven/NuGet proxy, etc.) so subsequent installs never leave your machine. Pull all required Docker images in advance and, if needed, save them with docker save so they can be transferred to an air-gapped machine.

4. Enable offline debugging

Configure your editor’s debugger to run against the local process or the process inside the Dev Container. VS Code’s built-in debuggers, Neovim with DAP (Debug Adapter Protocol), and JetBrains debuggers all work fully offline once language support is installed. For web front-ends, use the editor’s integrated browser or a local browser with device emulation. Run databases, caches, and other services as local Docker containers so integration tests and debugging sessions do not need external hosts.

5. Optional but powerful: local AI assistance

Install Ollama (or LM Studio) and pull a coding-focused model (for example a quantized Qwen or similar coder model) while you still have connectivity. Point an offline-capable AI extension or tool (Continue.dev, Cline, Aider, or a dedicated local IDE) at the local Ollama endpoint (http://localhost:11434). Once configured, autocomplete, chat, and agent features work without any external API calls.

6. Prepare offline documentation and search

Install a local documentation browser such as Zeal or Dash and download the docsets for the languages and frameworks you use. Keep a local copy of important project docs, READMEs, and API references. Use local search tools (ripgrep, or rga for PDFs) so you can find information without the web.

7. Day-to-day offline workflow

Clone or open the project, start any required Docker services, open the editor (or reopen in the Dev Container), and work. All builds, tests, and debugging sessions stay on the local machine. When you regain connectivity you can push changes, pull updates, and refresh caches as needed.

Minimal viable checklist before going offline

Editor + language tools + debugger installed and verified. Project dependencies installed and Docker images pulled. Any local package mirror or cache warmed. (Optional) Local LLM downloaded and connected to your AI tooling. Documentation offline and searchable.

This combination—host tools + Docker/Dev Containers + cached dependencies + optional local AI—gives you a reliable local-first environment that supports full coding and debugging even when the network is unavailable. Start with the editor and Docker setup for your primary language, then layer on the offline caches and AI components as needed.