I’ve been building onju, a tool that makes it incredibly natural to interact with computers Make conversational requests, powered by LLMs & the context of the app you’re in, without breaking your flow Here’s a first example of using it with... Terminal! (🧵 for access)

Transcription has to be fast & flawless for a good UX. To do this the model runs on-device and is conditioned on the app you’re in and with previous commands or messages to ensure accuracy with ambiguous words. (Recordings never need to leave your device)

Responses are currently using GPT4, chained when needed with memory, preferences & integrations. Although it can provide highly specific answers, it works across all apps w/o extensions needed for each & allows credentials for integrations to be centrally & securely managed

I believe the magic happens when you pair 🍐 voice (for rich requests & feedback) with physical inputs (for shortcuts & confirmation) to get the answer where you need it (More demos in different apps coming!) Sign up for access at

@onjuai brilliant! something like this running in-device with @ApacheTVM or similar is honestly the dream, can't wait to give it a spin tho

@onjuai This is awesome! Do you think you could achieve reasonably similar results with on-device open assistant, or replits copilot model?

@onjuai but did you really fix all the bugs?

@onjuai Haha, well... define "you". We're a pair...