██████  ██    ██ ███    ██    ██████  ███████ ██    ██
 ██   ██ ██    ██ ████   ██    ██   ██ ██      ██    ██
 ██████  ██    ██ ██ ██  ██    ██   ██ █████   ██    ██
 ██   ██ ██    ██ ██  ██ ██ ██ ██   ██ ██       ██  ██
 ██   ██  ██████  ██   ████ ██ ██████  ███████   ████

run.dev is a single Rust binary that replaces your entire local dev environment setup. Instead of juggling terminal tabs, nginx configs, docker-compose files, and manual /etc/hosts entries, you get one dashboard that manages all your services, gives them real HTTPS domains, and monitors everything. Think of it as "what if someone built the perfect local dev experience from scratch in 2026." You install it with one command, add your projects, and every service gets a pretty URL with SSL, live resource monitoring, and crash recovery — all from one TUI dashboard.

Docker Compose orchestrates containers. run.dev manages processes directly on your machine — no containers, no virtualization overhead, no Docker daemon eating 4GB of RAM in the background. Your services run natively, exactly like they would if you typed the command yourself. run.dev just handles everything around them: domains, SSL, reverse proxy, process monitoring, and crash recovery. If your workflow is "I run npm run dev and cargo run in a bunch of terminals," run.dev is for you. If you need containerized deployments, Docker is for you. They solve different problems.

Not yet. macOS and Linux only. WSL2 might work but isn't officially tested or supported. Windows native support is on the roadmap, but the /etc/hosts and port forwarding mechanics are fundamentally different on Windows and need their own implementation. If you're on Windows with WSL2, give it a shot and let me know how it goes.

Yes. The current version is fully open source under the MIT license — the code is on GitHub. No telemetry, no gating. I may introduce a Premium version down the line with advanced features, but the open-source core will always be free.

Anything that runs with a shell command. run.dev has smart detection for Node.js (package.json), Rust (Cargo.toml), Go (go.mod), Python/Django (manage.py), Ruby/Rails (Gemfile), and Docker Compose files. But you can add any service that starts with a command — PHP, Java, Elixir, whatever. If you can type it in a terminal, run.dev can manage it. It also auto-detects ports from command flags, .env files, and framework defaults.

run.dev generates trusted SSL certificates automatically using mkcert (if available) or its built-in pure-Rust certificate generator (rcgen). When you add a service, it creates a certificate for your domain, configures a reverse proxy with SNI-based routing, and everything just works — green padlock in the browser, no warnings, no manual trust steps. Certificates auto-renew before expiry. The whole thing is zero-config: add a service, get HTTPS. That's it.

No. The installer downloads a prebuilt binary for your platform. If prebuilt binaries aren't available for your architecture, the installer will automatically install Rust and build from source — you don't need to do anything manually. Once installed, run.dev is a single binary with zero runtime dependencies. No Node, no Python, no Docker, no nothing.

run.dev integrates with Claude Code — Anthropic's local CLI agent — because it's the only AI that runs entirely on your machine. Your code and logs never leave your system. Here's where it helps: when a service crashes, Claude reads the stderr and explains what actually went wrong instead of just showing you an exit code. Press / in the dashboard and ask anything about your running services — it knows your project structure, routes, and ports. For common errors like port conflicts or missing modules, it suggests a one-key fix. That said, run.dev works flawlessly without any AI. Every feature — the dashboard, domains, SSL, proxy, process management — is fully functional on its own. The AI is a layer on top, not a dependency. No API key needed. Disable anytime with --no-ai or just don't configure it.

The installer is fully transparent about everything it touches. It installs the binary to /usr/local/bin, adds a small hosts helper script with a sudoers rule (so it can update /etc/hosts without asking for your password every time), and sets up port forwarding rules (pfctl on macOS, iptables on Linux) to route ports 80 and 443 to the local proxy. All changes are scoped to localhost only — it never touches your outbound network traffic. The installer shows you the full list before proceeding and requires your explicit confirmation. If the installation fails partway through, it automatically rolls back everything it changed. All changes can be reversed with rundev uninstall.

No — press q and your services keep running in the background. The proxy stays active, your domains keep working, everything continues as-is. Next time you open run.dev, it reconnects to your running processes automatically via PID persistence. If you actually want to stop everything, press Q (capital Q) which triggers a graceful shutdown of all services, or run rundev down from the command line. This means you can use run.dev as a quick status check without disrupting anything.