What are logbooks?

Logbooks are a lightweight pattern for keeping unfinished work usable across agents, tools, and sessions.

At its core, a logbook is one table — or a small set of connected tables — with a stable entry contract. Entries can be appended, patched, and annotated as work evolves. Agents and humans query the structure through tools instead of rereading everything in context.

research-logbook/
├── findings.csv      # main entries
├── comments.csv      # optional annotations
└── actions.csv       # optional action history

Sometimes a logbook is even simpler:

id, title, status, owner, next_step, updated_at

How logbooks work

1. Write: agents and humans add or update structured entries as work progresses.
2. Query: tools answer common questions by filtering, sorting, aggregating, or joining fields.
3. Act: selected entries drive next steps such as review, reporting, tracker commits, or new agent runs.

Why they matter

Without a logbook, intermediate work tends to disappear into chat history, scattered docs, or half-finished tickets. With a logbook, work can carry forward: one pass leaves behind structured results, and the next pass builds on them. That makes logbooks especially useful for deep workflows, draft or staging workflows, and background agent work.

What logbooks are not

Key advantages

• Reusable: partial work survives across sessions.
• Queryable: common questions are answered through tools, not rereading.
• Flexible: start with a sheet or CSV, grow into JSONL or SQLite if needed.
• Staged: shape work before committing it to execution systems.

Next steps

• Read when to use a logbook
• Start with one small schema
• See examples
• Choose a backend