We’re the team behind Kortix, the open-source CompanyOS. It keeps a company’s agents, skills, memory and connectors as files in a git repo you own, works with any model, and runs on your servers or ours. On this blog we share what we’re learning from putting agents to work for real teams: in Slack and Teams, on background coding jobs, on a schedule. We also write a lot about keeping a human in charge of the steps that matter.
Handing an agent a Stripe key is a decision about where that key lives. The model's instructions are the least reliable part of the arrangement, since everything in a context window competes with them
Cowork on Linux runs in a QEMU/KVM virtual machine the app hosts on your own machine, so the hypervisor and the record matter more than the distro. You can install Claude Desktop on Linux today. Anth
Map AI agent workflow failure modes to concrete controls, with a reference implementation on open-source Kortix. Agent workflows fail in ways a generic retry does not fix. A tool reports success with
