Google Keep Extractor

Parse JSON files from Keep export and convert them to Markdown files. Attachments and images are preserved.

You can download Google Keep backup file by visiting:

https://takeout.google.com/takeout/custom/keep

You can export data to a .zip or .tgz archive which has the following structure:

takeout-<timestamp>.<zip/tgz>
└── Takeout
    └── Keep
        └── JSON & HTML files, and attachments

Copy Takeout folder into root of this repo, and run the script:

python google_keep_extractor.py

Python 3.10 or later is supported. No additional packages are required.

After running the script, Markdown files with timestamp in the filename are created and written to export folder.

Main features

• Exports notes to individual .md files
• Includes attachments
• Adds Keep's Labels into the files
• Pinned notes have their titles starting with [PINNED]
• Archived notes have their titles starting with [ARCHIVED]
• Trashed notes are skipped (note that they're included in the backup, but we explicitly skip them)

Inspect testing folder to see sample Keep backup files and output produced for it by this tool.

Use Cases

Obsidian

Export is a collection of Markdown files which can be easily imported into Obsidian vault.

Apple Notes

You can also export your Google Keep backup into Apple Notes. Notes supports RTFD (Rich Text Format Directory) imports of multiple files. Bear (available in the App Store) can be used to import the Markdown files into it, and export RTF files. This should be easier than tinkering with other tools like Pandoc and creating RTF files directly.

Development

Note

uv is required for justfile commands to work correctly.

Prerequisites:

• uv installed and on PATH
• just installed and on PATH

Installation and testing

Development dependencies are stored in requirements-dev.txt. tox is used to specify test envs and how to install dependencies.

To install package in the editable mode with dev deps:

just venv

To run pre-commit checks:

just pre-commit

To run tests against all supported Python versions:

just test

To measure code coverage:

just coverage

To run all the above tests & checks:

just all

Regenerate expected output after any changes in the extraction process:

just regenerate