Admin & diagnostics
Check what changed, whether your data is current, and whether the AI is actually connected — without asking anyone.
Automation you can't inspect is automation you shouldn't trust. The System section of the app exists so you can always answer three questions yourself: what changed, is my data current, and is the AI actually connected?
Activity — what changed and who did it
Every consequential action is recorded as an activity event: a source snapshot recorded, a blend saved, custodians anchored, a target generated. Each entry says what happened and which surface did it — you in the app, or your connected AI assistant.
That's the accountability half of letting an assistant do batch work on your behalf: it can record thirty things while you're asleep, and you can read exactly what it did in the morning.
What's New
A running log of what's shipped, in plain language, with unseen items counted in the header. The product changes often; this is where it tells you how, without a release-notes scavenger hunt.
Data freshness and sync health
Your portfolio sync reports its own status: when it last ran, whether its connection is still valid, and a direct path back into settings to refresh it if it's expired.
The same honesty runs through the rest of the product — sources carry their own age, and a source with no declared publishing cadence is never labeled fresh. Nothing here reports health it hasn't verified.
AI provider connectivity
If you've configured your own AI provider for the AI Council, the settings page shows whether it's actually reachable — a lightweight check by default, plus an opt-in deep check that round-trips a real request end-to-end.
So you know before a session whether an answer will come from your model or from the clearly-labeled simulated fallback.
Under the hood
Some tooling is scoped to the workspace owner: a browser over every stored record with its full version history, service status, and logs. Most people never need it; when something looks wrong, the raw record is there rather than hidden.
The health endpoint at /api/v1/health reports database connectivity, required
configuration, disk space and event-loop health — a fast, honest answer to "is it up" for
you or an uptime monitor.