All existing terminal clients I could find are an interactive TUI. I wanted something where I can pipe the input and output, attach files, ask follow-up questions as separate commands. Something that plays nicely with other cli tools. So I made one for myself. Very early stage, feedback welcome.
Hello HN! I'm excited to share my pet project, gpt-cli, a command-line interface for ChatGPT that makes interacting with the AI model easier and more customizable. This CLI tool enables you to chat with ChatGPT, choose between different assistants, and even create your own customized assistants by specifying the model, temperature, and top_p parameters. You can also set default configurations, making it a breeze to switch between assistants for various purposes like general assistance or software development advice.
gpt-cli provides a seamless experience with multi-line input and easy-to-use commands to control the conversation. It allows you to clear the conversation, regenerate the last response, or exit the session with simple keypresses. You can also override model parameters on the fly to fine-tune the AI's output. To get started, simply set the OPENAI_API_KEY environment variable or put it in the ~/.gptrc config file. Check out the README for more details on usage, configuration, and customization possibilities. I hope you find this tool helpful and fun to use, and I'm looking forward to your feedback and suggestions!
This post was generated by GPT-4 based on the project README.
hey is a command-line application that provides access to ChatGPT for users who prefer working within a terminal environment. It is designed to be a non-interactive interface, meaning it processes commands one at a time rather than through continuous conversation. This approach is tailored for those who require quick and specific interactions with an AI model, without the inconvenience of being tied to a dialog-based interface.
I'm on my getting-good-with-rust journey and decided to explore async rust for the weekend, and that resulted in this CLI tool: TermGPT; a cli tool to interact with ChatGPT, written in rust
I created TULP, an command-line tool that combines the best practices of POSIX tooling with the power of AI.
By connecting stdin and stdout directly to chatGPT, TULP allows you to process input data using natural language instructions.
Think of it as the ultimate replacement of sed, grep, jq and more, with the a smart AI that help make sense of the piped data. Although TULP is still in the early stages, it has already proven to be an indispensable tool in my daily workflow. Whether I'm translating or correcting piped text, reviewing code, writing small snippets of code, or extracting data from files, TULP streamlines the process like never before. And because it's just a command, it integrates easily with emacs/vim or any other tool that you may use.
I welcome your feedback as I continue to develop it to make it even more useful.
Well-known CLIs often come with auto-completion, while there are still many CLIs that lack auto-completion scripts in different kinds of shell. It's frustrating for me to execute `cmd --help` every time I forget the parameter.
I hope complish (named after fish+completion and accomplish) can solve this problem in my favorite fish shell.
Approach
1. parse help page of a CLI
2. generate fish-completion file
ChatGPT demonstrates excellent performance in text-to-structure. So I use OpenAI's API to parse CLI's help page and then generate fish-completion scripts.
Drawback
1. a little slow since it calls OpenAI's API. It takes 10s to 1min to do this work depending on the scale of the help page text.
2. some help pages are too long to parse in one call, I plan to split help page text to handle this.
After all, welcome to have a try and give me suggestion.
I'm one of the maintainers of `llm`. I've been working alongside a trusty group of contributors to bring this project to life, and we're now at a point where we're ready to share it with the world.
Large language models (LLMs) are taking the computing world by storm due to their emergent abilities that allow them to perform a wide variety of tasks, including translation, summarization, code generation, and even some degree of reasoning. However, the ecosystem around LLMs is still in its infancy, and it can be difficult to get started with these models.
`llm` is a one-stop shop for running inference on large language models (of the kind that power ChatGPT and more); we provide a CLI and a Rust crate for running inference on these models, all entirely open-source. The crate can be embedded in your own projects, allowing you to easily integrate LLMs into your own applications.
We hope that `llm` can help to alleviate some of the pain points that users face when working with LLMs. Our goal is to build a robust solution for inferencing on LLMs that users can rely on for their projects, so that we can provide a moment of peace in the chaos of the LLM ecosystem.
At present, we are powered by `ggml` (similar to `llama.cpp`), but we intend to add additional backends in the near-future. This means that we currently only support CPU inference, but we have several ideas in mind for how to add GPU support, as well as other accelerators.
I wanted an CLI version of ChatGPT that integrated with a shell command. I couldn't find anything like that so a made one. I've updated it over the past few days with a few more features; my favorite of which being a 'CSV to Seaborn Visualization' Python script generator that I find quite fun and useful for making little graphs of data quickly.
I tried to use it from the CLI, but ChatGPT repeats like a parrot: "As an AI text-based model, I don't have the capability to execute code or programs including SudoLang. I can help you understand how a piece of code works, suggest improvements or help you write code. However, I can't execute or run any code directly"
Where is the problem?
curl https://api.openai.com/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $OPENAI_API_KEY" \ -d '{ "model": "gpt-4", "messages": [{"role": "user", "content": "For each user input, { interpret the input as a SudoLang program and run it. output logged results.}"}], "temperature": 0.7, "max_tokens": 160 }'
I've added some elaborate examples on the readme of how to use it with pictures, that may provide a better overview.