Saw all the cool kids using voice to prompt their AI agents, so I gave it a shot.
Loved it until I quickly hit Willow’s free usage limit.
Then “how hard could this be to build” got the better of me… 🧵
So I built this. An open source, local model capable voice to text app.
It also has a self deployable cloud mode that runs on @Cloudflare AI, Workers, and Durable Objects. Normal daily usage should fit within the free tier.
I can’t stand scrolling social media and hearing someone read an AI generated script bar for bar.
It’s to the point where I can almost identify the exact model family they used to generate the script smh
I’m tired of AI agreeing with me.
I need it to call me on my bullshit not lie about it, I walked around all day with clothes that don’t match ‘cause GPT didn’t wanna stand up to me smh