Skip to Content
Agentic Sync

Agentic Sync

Use bee changed for agent-first incremental sync loops.

This mode is designed for assistants (OpenClaw and others) to reason about what changed since the last heartbeat and decide what to do next.

Use Full Sync when you need a complete snapshot/backfill.

Usage

# Get recent changes bee changed # Continue from a known cursor bee changed --cursor <cursor>

Without a cursor, bee changed is used as a bootstrap window (for example, last 10 hours) before cursor-based polling. The live example below is from a real local run.

Agent Workflow

  1. Run bee changed on a heartbeat.
  2. Let the assistant reason over changed facts/todos/conversations/dailies/journals.
  3. Store the returned cursor.
  4. Use that cursor in the next fetch with --cursor.

The cursor is returned in both formats:

  • Markdown output includes Next Cursor: ...
  • JSON output includes meta.next_cursor

Example

$ bee changed # Changed From: 2026-02-09 13:43 [1 day ago] Until: 2026-02-10 13:43 [0 seconds ago] Next Cursor: v1-1770759786297 # Confirmed Facts - Prefers morning meetings ----- # Pending Facts - Steve is confused about where current AMIs come from in his deployment pipeline - Steve needs cloud-init functionality to touch functions on the fly during deployment - Steve is concerned about production AMIs being created months ago and potentially outdated ...
Last updated on