> ## Documentation Index
> Fetch the complete documentation index at: https://docs.meshpilot.app/llms.txt
> Use this file to discover all available pages before exploring further.
# Brain
> Learned patterns, confidence tiers, and the reflective surface.
The Brain is the cockpit's reflective surface. It shows the patterns your agents have learned — what worked, what didn't, what's been corroborated across enough evidence to act on confidently.
It's not a generic analytics page. It's the answer to *"why is the agent proposing this?"*
## What's on the page
* **Strong / moderate / weak patterns** — badges or a micro-strip at the top showing how the brain is splitting its current knowledge.
* **Patterns table** — each learned pattern with a plain-English summary (e.g. *"calls placed before 10am IST have \~2× higher pickup rate"*), the confidence tier, and the evidence count.
* **Confidence view** — drill into how a pattern's confidence has moved over time as new evidence corroborates or contradicts it.
* **Evidence / corroboration signals** — the anonymised brand count, the operator-label history (followed / overrode), and the timestamps.
* **Reflector explanation footer** — a short note on how patterns get promoted from observation to actionable wisdom.
## How it relates to your agents
Every agent emits observations as it works. When the same pattern is seen across enough brands, the brain's reflector promotes it into anonymised cross-brand wisdom. Agents recall patterns before drafting proposals; you can label whether a pattern's advice actually worked, which closes the confidence loop.
Patterns are anonymised before they ever leave your workspace — see **Brain & memory → Privacy** for the invariants.
See what your agents have learned.