
nlsh is a terminal interface that translates plain English into shell commands. Instead of memorizing obscure flags and syntax, you simply type what you want to do—and nlsh converts your natural language into the appropriate bash command. It runs locally on macOS or Linux, requires Python 3.8+, and has already been installed over 9,300 times.
Type commands like "list all python files" or "kill process running on port 3000" and nlsh outputs the correct shell command. It handles everything from file operations to git commits and system management.
nlsh includes built-in commands like !api to securely change your LLM provider API key (currently supports Gemini), !help to view the help menu, and !cmd to run raw shell commands directly without natural language processing.
Your prompts are only sent to the AI provider you choose—nlsh does not store your command history. The entire codebase is 100% open source, giving you full visibility into how your data is handled.
"Stop memorizing flags. Just type what you want."
This philosophy captures nlsh's core value: it removes the friction of recalling exact command syntax. Instead of pausing to remember whether it's -rf or -fr, you describe the task and let the tool handle the translation. The open-source nature and local-first design also mean you're not locked into a proprietary service or forced to share your terminal activity.
You spend significant time in the terminal and want to reduce context-switching between your work and documentation. It's especially useful if you manage multiple projects, work with unfamiliar tools, or simply prefer describing what you need rather than memorizing command structures.
Other tools you might consider
DiffSense uses the native AFM 3B model on Apple Silicon to generate git commit messages for free. It runs locally with zero latency, ensuring your code stays private. Features customizable message styles and powerful alias macros.
Blueberry is a Mac app that combines your editor, terminal, and browser in one workspace. Connect Claude, Codex, or any model and it sees everything.
Ask questions about any repo and AI agents will clone, explore, and grep the source code to provide source-backed answers.
AI can write code brilliantly but debugs blindly. It can't see your app, logs, or what users did, so you waste time explaining. BetterBugs MCP gives AI complete context to fix the bugs instantly.
Loading comments…
Maker
async_apple
Project Info
Product Keywords