Logseq Page CLI

A tiny Python CLI that reads Logseq's SQLite graph (db.sqlite), decodes the Transit data inside, and prints any page (regular or journal) as a plain-text outline. It also resolves raw UUID links ([[692cf4c0-…]]) to their human-friendly titles.

Features

• Works with the current Logseq db.sqlite format (Transit-encoded datoms stored in the kvs table).
• Accepts either a page name (Travels, Library, Peru) or a journal date (2025-12-03).
• Defaults to today's journal page when no argument is passed.
• Replaces UUID-based links with the referenced block/page title for easier reading.

Requirements

• Python 3.11+ (any version supported by transit-python).
• transit-python==0.8.302 (already declared in requirements.txt).
• Logseq graph folder containing db.sqlite, db.sqlite-shm, and db.sqlite-wal (copy/symlink the SQLite files next to the CLI).

Setup

# create / reuse a virtualenv with uv (recommended)
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt

If you prefer vanilla pip:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Usage

# default: today's journal page
uv run logseq_page_cli.py

# specific journal date
uv run logseq_page_cli.py 2025-12-03

# regular page or tag
uv run logseq_page_cli.py Travels
uv run logseq_page_cli.py "Library"

The script prints a header such as Journal page: 2025-12-03 [entity 16815] followed by an indented list of blocks. If the entire page is empty, you’ll see (no blocks recorded for this page).

Notes

• The CLI is read-only; it never writes to the database.
• Because Logseq stores every “fact” as a datom, the script rebuilds all entities in memory. For very large graphs this can take a few seconds.
• If you add new notes and do not see them in the output, ensure Logseq has synced changes to db.sqlite (e.g., by closing Logseq or forcing a save).
• This script is created by Codex CLI.

License

MIT. Do whatever makes your graph happier.