AIChat (https://github.com/sigoden/aichat) nearly fulfills all your requirements.

The only difference is that you provide subcommands, while aichat requires users to create various roles.

t2x 'Write a ...' => aichat 'Write a ...'

t2x ask file.md 'What is...' => aichat -f file.md 'What is...'

t2x summary /path/to/file => aichat -r summary -f /path/to/file

t2x summary "This is a..." => aichat -r summary "This is a..."

t2x orc /path/to/image.png => aichat -r orc -f /path/to/image.png

> Note, In AIChat, `-r/--role` specifies the role to use, `-f/--file` indicates the file or URL input.

oooo neat! I hadn't heard of aichat before, I'll definitely take a look.

Any plans to integrate with https://openrouter.ai ?

that would be sweet

I'm curious, what specifically you're interested in from openrouter? Do they offer specific LLMs, or is it just a great way to kick the tires on new models as they emerge?

The original intention behind t2x was to be "functionality first", and make an opinionated call on which LLMs get used for specific functionality. However I think the way to go will be sane defaults and then allow user-specified models from something like openrouter.

hi friends!

I'm the author of t2x and shrugginface.com, I'm surprised to see this landed on HN a month after I first shared it.

I originally planned to make it real during the holiday break but got distracted by other projects. This is the inspiration bump I needed, so I'll get back to shipping over the coming weeks

Hope it’s okay I shared it here. Saw it in my RSS reader and figured HN would enjoy it.

Glad to hear it served as inspiration. The video looks promising!

I wish someone would do something like this but without the preamble like "t2x ask 'bla bla'"

All I want to do is talk to my computer, like:

"Sort these files" or whatever. Straight away. No noise.

And it should have a history of what was done already, so it knows what I'm talking about.

If people are concerned about safety, it could just be a 'mode' or an app that isn't a terminal but looks like a terminal.

> "Sort these files"

There's aichat, which you can use in REPL mode to do exactly that: https://github.com/sigoden/aichat

Shameless plug: I mostly use my own work: https://github.com/codespin-ai/codespin

But it requires you to pipe: `ls | codespin go sort`. Maybe I should add what you're suggesting.

If you run `ls -la | t2x "sort these files alphabetically"`, it will do the right thing.

However, subcommand like `t2x ask` are used to route you to a different model with different behavior. The ask subcommand currently makes requests to perplexity, where you can ask questions and get near realtime grounding from the world.

Why do you want this? Not only is it unreliable and inefficient to do such things with LLMs, but we already have easy ways to do them with existing tools.

I work with computers and do data wrangling on a daily basis. What am I missing?

I use a cli tool 100x/day for misc things; can hit openai or a local llm.

https://github.com/TheR1D/shell_gpt

Has nice shell integration.

heh, I wrote my own dotfile commandline tooling to do most of this already, I wonder if others have done the same

Mine's a Python CLI tool called LLM - https://llm.datasette.io/ - it uses plugins to provide support for 100+ different models: https://llm.datasette.io/en/stable/plugins/directory.html

Simon's LLM tool is probably the best of any of these sorts of things! So thank you for making it, I still use it regularly to kick the tires on new models, and it served as a significant inspiration for the design of t2x.

The idea behind t2x was to make a stripped-down cli with a "functionality first" mindset and remove the actual selection of specific models from the user's flow.