Txt: A fast, keyboard-driven terminal text editor for engineers

jlundberg 22 points 27 comments September 11, 2026
txt.hellman.io · View on Hacker News

Discussion Highlights (16 comments)

WesolyKubeczek

Since https://github.com/micro-editor/MICRO exists for some years now, a comparison would be welcome.

_ix

While they play up the keyboard-driven functionality, they showcase a pointer-driven workflow in the demonstration gif. Curious choice, no?

SrslyJosh

FYI: Claude has committed to the repo.

marjancek

> curl -fsSL https://raw.githubusercontent.com/ErikHellman/txt/main/insta... | sh With the amount of new seemingly well polished personal projects being built (thanks to the help of LLMs) I feel increasingly concerned by how easy it is becoming to get a worm/virus/troyan to a lot of people.

__fst__

Since going through the hassle of learning vim I can't use anything else anymore. And for the use case of "I built txt because I spend most of my time in the terminal, and switching to a different tool just to make a quick edit always felt like friction I did not want. The editors that live in the terminal either have steep learning curves built around modal paradigms, or they are too limited to be genuinely useful." it's literally only a handful of vim commands and shortcuts you need.

smithza

I live in emacs over ssh connections. I never use the GUI, just TUI version (so `emacs -nw`) and love it. You can edit files over ssh hops or docker hops using tramp. I don't believe there is a tramp equiv. in other terminal tools. VS Code over SSH is much too heavy handed for quick single file edits but emacs handles it gracefully.

rckt

I can understand startups relying on AI tools to test out the ideas. But personal/hobby projects? Feels weird. This is not the place to go fast. Take your time, enjoy the process, learn/make things.

rvz

I can tell a README which is a long as that is not written by the author, but of course by Claude. So, I ain't reading all of that.

vouaobrasil

I miss the days when people would release software not touched by AI. And I wish at least there'd be a disclaimer saying that this was AI-assisted, so I could avoid it.

messh

This is gonna be a hard name to google :)

hmartin

Is this even still under development? Last non-renovate commit was early June

cs1996

reminds me of https://zed.dev/ ?

david38

A comparison with emacs and vim would be nice.

sgarland

> The editors that live in the terminal either have steep learning curves built around modal paradigms, or they are too limited to be genuinely useful. > txt is not trying to replace your main IDE. It is the editor you reach for when you just need to edit something These two are in conflict. If you just want to edit something, and don’t know [n]vi[m] or emacs, then use nano. If you want line numbers, code folding, syntax highlighting, etc. then it sounds like you should spend the time to learn a more powerful editor. Honestly, vi isn’t that hard to learn the absolute basics, and you should spend the 15 minutes to do so, because it’s practically guaranteed to be available everywhere. :wq and i are about all you need to “just edit something.”

eviks

> GUI editors are the opposite problem — they are powerful, but reaching for the mouse to edit a config file or fix a typo breaks your flow. You don't need to reach for the mouse, are you not aware you can launch, switch apps, and type just with your keyboard? But also, how is your demo using a mouse of your whole selling point is anti-mouse?

vivzkestrel

- one of these days... - it would be absolutely amazing if one of the guys writing posts along the lines "here is our super duper text editor"... - would actually explain in a series of blog posts how they built a text editor from scratch and I am not talking about code... - I am talking about architecture, methodology, how each feature is implemented on paper before coding it, thought process behind one.. - I am amazed that we live in a world where people create all types of software and yet little to no blog posts ever discuss the thought process and methodology behind them... - like why did you do this... - why is this feature implemented in this way and what are the alternative implementations and why did you pick this implementation over the others

Semantic search powered by Rivestack pgvector
6,278 stories · 57,251 chunks indexed