Operate

Running the platform: who may sign in, which model it uses, what it may spend, and how it is deployed.

For the day-to-day screens, see Dashboard.

Installing for the first time? Read the first three pages in order: what the image needs before it starts, how to register the first sign-in method and administrators, then how sign-in is configured.

Page	Covers
Container image	What each process needs before it starts, the two database connections, and the order to run them
First-run bootstrap	Registering the first installation sign-in method and the first platform administrators
Authentication	Signing in, and who is allowed to
Triage engine	Choosing the model, capping what it may spend, and reading what it cost
Platform settings	Every tunable value, its range, and when a change takes effect
Slack	What to expect from Slack once it is connected
Network policy	Which addresses the platform may reach, and which it refuses

Two boundaries worth knowing before you deploy

Organisations cannot see each other. Every record carries the organisation that owns it, and the database enforces the separation rather than the application remembering to. Two organisations on one deployment share infrastructure and nothing else.

Credentials never leave in plaintext. Connector and Slack credentials are encrypted at rest and decrypted only at the moment of use. Nothing displays them back, and nothing writes them to a log.