My personal multi agent dashboard - mrum_

[buoyant brook]

👏 are these multiple claws running or all on one instance? interested in the steps to get something like this running, if open to sharing

[thin locust]

Would you share this design? Love it

[tawdry plank]

Krlecarp
yes of course, i will walk through the steps.
in this framework, the backend runs an orchestrator that can launch multiple specialised agents in parallel, but it limits active calls to two at a time. so the system stays stable and predictable under load.

all agents share the same openclaw_base_url, meaning they use one model service endpoint while still behaving as seperate roles. with the screen, its (main, group, research, dev) through different promts, memory context and model routing settings

Magnus
ofc, this design is a multi agent operations board. Each agent has a role, and the board shows their workload and health in one place, and it is colour coded too, so that i can easily track everything that's going on.

[thorn wren]

did you share this ability/interface somewhere？clawhub skill or github？anywhere i can get a hold of this