LangCommand is a revolutionary local inference command-line tool designed to seamlessly convert natural language descriptions into executable shell commands. Powered by the advanced llama.cpp framework, LangCommand simplifies command generation, allowing users to describe their needs in straightforward language. Here are some of its compelling features:

• Natural Language Command Generation: Transform your text descriptions directly into shell commands without any hassle.
• Customizable Models: Choose from a selection of preconfigured models or supply your own to create tailored workflows that fit your specific needs.
• Multiple Modes:
  • Loop Mode: Executes commands continuously until terminated.
  • Exit Mode: Executes a single command and exits back to the command line.
• Command Explanation: Get an optional breakdown of the generated command for better understanding and transparency.
• Tracing: Activate model tracing for debugging assistance and to monitor the execution process.

Usage

Using LangCommand is straightforward. Here’s how you can provide a natural language description:

    lac "your prompt"

You can also explore additional options by running lac without any arguments:

--------------------------------- LangCommand params ----------------------------------
-h,    --help, --usage                  Print LangCommand usage
--setup                                 Set up your LangCommand model: choose or customize
--show-args                             Show arguments you saved
--no-explanation                        Disable command explanation
--mode {loop,exit}                      Select the mode of operation.
                          - loop: Continues to choose and execute commands indefinitely.
                          - exit: Executes a single command and then stops the program.
--model-help, --model-usage             Print LangCommand default model arguments
--trace                                 Enable tracing for the execution of the default model

Supported Models

LangCommand supports a variety of robust preconfigured models. Check out the following options:

• Llama-3.2-3B-Instruct-Q8_0
• Llama-3.2-1B-Instruct-Q8_0
• qwen2.5-7b-instruct-q8_0
• codellama-13b.Q8_0

Feel free to provide your own model along with custom system prompts for unique implementations.

Troubleshooting

Experiencing memory issues? You might encounter the following error when loading models:

ggml_metal_graph_compute: command buffer 1 failed with status 5
error: Insufficient Memory (00000008:kIOGPUCommandBufferCallbackErrorOutOfMemory)

This often indicates that your system lacks sufficient memory for the selected model. Consider opting for a smaller model to mitigate this issue.

Community & Contact

Join the LangCommand community to connect with other users, share your feedback, and receive support:

• Join the LangCommand Discord group
• For any inquiries, reach out via email: Email me

LangCommand empowers users to streamline their command line tasks through intuitive natural language processing, making it an essential tool for developers and tech enthusiasts alike.