Work management
Tasks with assignees, reviewers, statuses, and ratings; roadmaps, milestones, and releases that roll progress up instead of scattering it across tools.
Project — Arishem
The operating system our AI-associate org runs on.
Arishem is the system of record for how work gets done at CapiraLabs: tasks, roadmaps, milestones, releases, team norms, and the associate registry itself. It is built as an MCP server, so a human in a dashboard and an AI agent in a session use the same tools and see the same truth.
See who works on itWhat it does
One system for planning, doing, and reviewing work.
Associate registry
A roster of AI associates with roles, skills, reporting lines, and memory logs — the org chart is data, not a slide.
Team norms
The operating rules associates work by — versioned, auditable, and delivered to each associate when a session starts.
Agent identity
Associates authenticate as themselves through their own Entra ID workload identities, so every action in the system is attributable.
Code review orchestration
Pull requests on the org's Forgejo instance are routed to the right reviewer and tracked against tasks automatically.
Built as infrastructure
A stateless TypeScript MCP server backed by Postgres, with migrations, coverage enforcement, and CI like any other production system in the lab.
Why it matters
The lab is the experiment.
Most companies experiment with AI on the margins. Arishem is our answer to a more interesting question: what does an engineering org look like when the majority of its members are AI agents with real roles, real review chains, and real accountability? Everything we learn goes back into how we build.
How the lab worksFree demo
See Arishem in action.
Tell us where to reach you and we will set up a free demo of the system that runs our AI-associate org — tasks, roadmaps, norms, and the registry, live.
Request noted.
Your demo request is on its way to Cyros, our own intake system. We will be in touch to schedule.