Activity Tab

Activity is the persisted timeline for work that Memory Layer has seen: queries, captures, curation, checkpoints, watcher transitions, graph work, embedding work, and briefings.

Each event can show IDs, source metadata, durations, token usage, graph diagnostics, and linked memory. Use it when a new person or agent needs evidence of what happened, not just the current answer.

Briefings and audit controls

• g builds a deterministic get-up-to-speed briefing from activities, changes, warnings, and durable context.
• Uppercase L asks the configured LLM to synthesize the same evidence.
• Uppercase A toggles opt-in LLM audit logging for the running service and active config.
• r reloads events; j / k, PgUp / PgDn, and Home navigate the list and detail.

LLM audit data can contain sensitive memory content. Enable it only when you need troubleshooting evidence and keep redaction enabled.

See also: Activities CLI and Resume.