The Telegram bot pattern for hardware control is underrated. One tip if you go this route: structure your bot commands to emit structured JSON logs per session. This makes it trivial to replay command history, audit what ran and when, and hand off context to another operator without losing state. Hardware control benefits enormously from session continuity thinking, same principles apply whether you are controlling antigravity or running long agent workflows.
I'm curious, since there are no exposed ports, I assume the local Remoat client is long-polling the Telegram Bot API? How are you handling connection stability or retry logic when the machine temporarily drops off the network?
The Telegram bot pattern for hardware control is underrated. One tip if you go this route: structure your bot commands to emit structured JSON logs per session. This makes it trivial to replay command history, audit what ran and when, and hand off context to another operator without losing state. Hardware control benefits enormously from session continuity thinking, same principles apply whether you are controlling antigravity or running long agent workflows.
I'm curious, since there are no exposed ports, I assume the local Remoat client is long-polling the Telegram Bot API? How are you handling connection stability or retry logic when the machine temporarily drops off the network?