Live Reasoning
When you send a request, the agents’ thinking steps stream in as they happen. You can follow the plan they form, the tools they decide to call, the status of each call, and how long each one takes — turning what would otherwise be a black box into a transparent, auditable trail. This visibility means you always know why an agent did something, not just what it returned — useful for reviewing an investigation, catching a wrong turn early, or handing the reasoning to a colleague.Operational Panels
Some tool calls are interactive. When an agent needs a richer surface than a chat message, it opens a panel — either inline in the conversation or as a tab in the Workspace — that you can watch update in real time.Live SSH terminal
When an agent runs commands on a host, a live terminal streams the remote
session so you can see exactly what’s being executed and the output it
returns.
Audit dashboards
An audit produces a self-contained, interactive dashboard rendered right in
the conversation — charts, findings, and drill-downs without leaving chat.
Plans & todos
For multi-step work, agents lay out a plan as a checklist and tick items off
as they complete them, so you can track progress at a glance.
Live report authoring
When Scribe writes a report, you watch it appear section by section in the
document editor — read-only while it’s being written, then yours to refine.
These panels are live views of work in progress. Consequential actions still
follow Myrmex’s human-in-the-loop model — generated automations, change
requests, and risk acceptances are created as drafts for you to approve. See
AI proposes, you approve.
Where to Go Next
Chatting with the AI
Direct the agents with
@-mentions, templates, and attachments.Tabs & split view
See where panels and results open, and work with two panes at once.