This is oddly specific to me I’m sure, but I drive alot for work. So I would like to request a greater stability for the connection that the mobile Replit app depends on.
The real feature that I request is the ability to verbally interract with the Agent. That way when I am driving or cooking, I can still be working.
Future expansion on this idea is when we have a team of people in the office, and we are discussing features and such, that the Agent can listen in, adjust notes and it’s task list and implement if agreed on in the meeting.
Raymmar, Manny, Mark, & Allen63 (on this forum) all suggested whispr flow (https://wisprflow.ai/).
Highly recommended. Also, try using the speech to text functionality. The agent can’t speak back, and it’s not a conversation like when chatting with GPT verbally. However, if you’re on a small device screen (mobile) and want to at least have a one way verbal conversation that works very well.
You can also try, Natural Reader (https://www.naturalreaders.com/) to have it read to you. Although, I agree a verbal back and forth between the agent and developer would probably make me cry.
For me, it’s a must have for accessibility development tools. For visual impairments, severe ADHD, Dyslexia.
I’ve got a project in the works to read code snippets aloud integrating natural readers paid tier API. However, I’d have to charge for the API out call usage if I open it publicly, because it’s expensive. Even with documented disabilities I only get a discount. I couldn’t eat those costs.
IMHO to date there is not a solid verbal code reader. I mean it’s awful. It makes taking tests with the tools nearly impossible. Especially if your school lacks adequate resources.
I also want it to go both ways. For example, try verbally speaking the code you want to type while using speech to text.
I’ll wait… jk I won’t because you’ll never come back.
Yet, every single day someone with disabilities is out there trying to make it work.
I’m only in the prompt creation phase of this project, and it’s on the end of a list. I want to craft/ test the prompt so that the tool works “out of the box” for users to re-create (quickly), add their own e-reader / LLM for verbal conversation.
It doesn’t exactly solve this niche problem but it might make it easier or more manageable.
Let me know if that sounds like something you’d use this and I’ll tag you when I share.
But even now, I’m at a BBQ and I’d rather be working on my project.
Note: I am 100% not a coder or programmer. So I use replit like a dumb-dumb. I tell it to build a feature, then ask it to fix and adjust things and it works. When the company is big enough, we will hire proper coders, but for now this works excellently.
Now, to be able to talk into the phone for the project instead of having a laptop open in the truck and struggling with the sun/glare and the keyboard, this would be way easier and faster for me.
Especially because I just found the service for the multitasking.
I would love to check out your stuff when youre ready!
My guy lol talking to it (speech to text) it’s the same process as speaking a text message.
I presume if you are connected to your Car Bluetooth you could have your phone on the mount, and then speak the changes into text and send it. In fact, I’ve done this…
It just can’t speak back to you unless you copy what it says and paste it into a reader, or have your phone read it aloud. However, if you have it read code aloud you’ll notice it’s not good. It’s not good at all lol
lol talking to it (speech to text) it’s the same process as speaking a text message.