Hi! I am trying to recreate the glass tablet the mc uses for writing in Passengers (2016). It’s a tablet where you dictate, and you point and talk for editing like it’s a scribe (no keyboard etc).
I always get tired from the mechanical work when writing essays for class or other personal writings. I tried dictating (WisprFlow etc) but there’s still the work of splicing sentences together and maintaining punctuation and rewording that I find tedious, so I am making a tool to minimize that friction.
In my project you leave the mic on, and while you talk an LLM figures out if you want to write it down or if you are asking for something to be changed. I used it for a couple personal blog posts so far and I found it nice for editing because when I want to change the structure of a sentence or something I just highlight and talk over it instead rather than putting my hands to the keyboard. I didn’t like Gemini in Google docs or chat bc it just feels like you’re fighting the LLM, and just dictation tools feel optimized for one-shot content.
I thought it might be useful for others as well (writing essays or blogs or journals etc). Please try it out and give honest feedback!
It’s just a prototype so there’s still some issues:
- the LLM can get stuck in reasoning loops which makes edits fail sometimes
- ambiguous words / instructions can fail in reasoning loops
Some ideas for what I would want to add/change
- Let you drag text around, and when you drop it the surrounding text gets updated for spacing and flow automatically so that you can just rearrange sentences.
Hi! I am trying to recreate the glass tablet the mc uses for writing in Passengers (2016). It’s a tablet where you dictate, and you point and talk for editing like it’s a scribe (no keyboard etc).
I always get tired from the mechanical work when writing essays for class or other personal writings. I tried dictating (WisprFlow etc) but there’s still the work of splicing sentences together and maintaining punctuation and rewording that I find tedious, so I am making a tool to minimize that friction.
In my project you leave the mic on, and while you talk an LLM figures out if you want to write it down or if you are asking for something to be changed. I used it for a couple personal blog posts so far and I found it nice for editing because when I want to change the structure of a sentence or something I just highlight and talk over it instead rather than putting my hands to the keyboard. I didn’t like Gemini in Google docs or chat bc it just feels like you’re fighting the LLM, and just dictation tools feel optimized for one-shot content.
I thought it might be useful for others as well (writing essays or blogs or journals etc). Please try it out and give honest feedback!
It’s just a prototype so there’s still some issues:
- the LLM can get stuck in reasoning loops which makes edits fail sometimes
- ambiguous words / instructions can fail in reasoning loops
Some ideas for what I would want to add/change
- Let you drag text around, and when you drop it the surrounding text gets updated for spacing and flow automatically so that you can just rearrange sentences.