A telegram bot experience to control a machine via computer-use-preview!

Community
1

I built a fully functional telegram bot experience to control a virtual machine using the new computer-use-preview API. It’s really fun to mess around with. I just had it go through all the steps to order me a Dominos pizza on its own, among other things.

Thinking about open sourcing the code, if there’s interest here. Let me know!

A chatbot named "Computer Use" is executing tasks to search for the latest news about OpenAI in a messaging interface. (Captioned by AI)
A chatbot named "Computer Use" is executing tasks to search for the latest news about OpenAI in a messaging interface. (Captioned by AI)1290×2796 209 KB

2

And here’s an example of some of the supported commands

A smartphone screen displays a messaging app interface with a list of bot commands for managing automation tasks. (Captioned by AI)
A smartphone screen displays a messaging app interface with a list of bot commands for managing automation tasks. (Captioned by AI)1290×2796 233 KB

3

Nice! Thanks for sharing.

I moved you to Community and added a project tag. If you could keep your updates here, it would help everyone keep up to date.

As we’re a developer community, can you share with us some of the pitfalls you ran into while coding this?