Skip to content

Overview

The Operations page is your control center for monitoring and maintaining your project's AI memory system. Navigate to Operations to access all operational tools.

Tabs

The Operations page has five tabs:

TabPurpose
On-CallReal-time monitoring of system status, performance metrics, and operational health
SessionsCoding sessions captured by the VS Code extension: commits, file changes, and enrichment pipeline status
ReviewMemory items flagged by staleness detection — accept, edit, or archive items that may be outdated
ProblemsIssues and errors affecting your AI system performance
HygieneAI-powered cleanup agent for renaming, merging, and deprecating memory items

On-Call

The On-Call tab provides a health overview of your project's memory system, including recent session activity, enrichment job status, and error rates.

Sessions

Browse all coding sessions captured by the VS Code extension. Each session card shows the source, status, event count, and number of extracted memory items. Click a session to view its detail view with the pipeline timeline, events, and jobs.

Sessions are created automatically when the VS Code extension captures git commits. An active session is reused for up to 4 hours before a new one is created.

Info

Sessions can also be accessed from the Memory page's Sessions tab. Both views show the same data.

Review

The Review tab surfaces memory items that need human attention. Items appear here for two reasons:

Staleness detection

When new commits modify files referenced by existing memory items, those items are automatically flagged as review. This happens during the post-enrichment staleness check — if 2 or more of an item's referenced files were modified in a recent commit, the item may be outdated.

Low confidence

Items with a confidence score below 50% may contain inaccurate information and benefit from human review.

Sub-filters

Toggle between two views:

  • Needs Review — Items flagged by staleness detection (status = review)
  • Low Confidence — Items with confidence below 50%

Actions

ActionEffect
AcceptSets status to active and resolves any associated problem docs
EditOpens an inline editor to modify the item's title, facts, and metadata. After saving, the item is also accepted
ArchiveSets status to archived and resolves associated problem docs
Boost(Low Confidence only) Sets the item's confidence to 80%

Bulk operations

Select multiple items using checkboxes, then use Accept All or Archive All to triage in bulk.

Problems

The Problems tab tracks issues detected by the system, including enrichment failures, staleness warnings, and contradictions between memory items.

Hygiene

The Hygiene tab provides an AI-powered cleanup agent that analyzes your memory items and proposes safe, minimal edits:

  • Rename titles for clarity
  • Retag items with better categories
  • Merge duplicate items
  • Deprecate superseded items
  • Flag quality issues

The agent runs in two phases: first it analyzes and proposes actions, then you review and selectively apply them. No changes are made without your approval.

Next steps