cc-bridge
Run and manage my coding-agent sessions from anywhere, in the browser.
✎ poke an agent — all faked
It started on Telegram. I'd piped my coding agents through a bot so I could poke them from my phone, away from the desk. It worked, but I wanted to actually see them, so I built the web version. I never stopped using it.
By now I'm running a whole bunch of these agents — roughly one per project, sometimes a dozen at once — and cc-bridge is where I conduct them. One glance tells me which ones need me; I drop into any one, type or talk to it, and watch it work. It's the cockpit I run the studio from.
This is the cockpit I run the studio from.
Where you come in
Orchestrating agents is the core of how everything else here gets built — cc-bridge is me building the controls for that, to my taste. It's the reason one person can move like a team, and that pace is what I bring to anything we build together.
If we build together
When we build something together, I set you up with your own agents and this is where you'd run them. You give one a task in plain language and watch it work — its steps stream by live, then the result lands. You can run several at once, switch between them, start or stop them. It's a control room for a small crew, and it's real today.
I teach you the rest by doing — how to direct an agent, correct it, break a big goal into steps it can run. You bring the taste; the agents carry the grind. You don't need to have built anything before, and knowing how to get real work out of AI is yours to keep.
Where it's going
I'm turning it into something other people can run their own agents through, so it stops being only mine.