Show HN: Prompt-run – run .prompt files against any LLM from the terminal (github.com)

A new command-line interface tool called "prompt-run" has been announced, enabling users to run `.prompt` files directly against any large language model (LLM) from their terminal. This tool treats prompts as code, allowing for a more organized and efficient approach to prompt management, which often gets cluttered across various platforms. Notably, prompt-run operates entirely on the user's machine without any backend, telemetry, or cloud components, ensuring enhanced privacy and security. Users can install it easily via pip and execute prompts with straightforward commands. The significance of prompt-run lies in its ability to streamline the deployment and testing of prompts, which are essential for LLM interactions. By allowing prompts to be committed alongside code in version control systems, it facilitates easier reviews, comparisons, and modifications. The tool features a simple .prompt file format with YAML frontmatter for metadata, enabling users to declare variables, set model parameters, and validate file integrity without necessitating complex frameworks. Additionally, it supports various operations like streaming responses, dry runs, and comparing outputs, distinguishing itself from existing tools like LangChain and promptfoo by offering a focused, no-frills solution for LLM interactions.