google-takeout-utils

CLI tools for searching, reading, and extracting emails from Google Takeout mbox exports. No conversion or import into an email client needed.

Quick start

No install needed — run directly from PyPI with uvx:

cd /path/to/your/takeout-folder
uvx google-takeout-utils@latest search-email --help

The @latest suffix ensures uvx always fetches the most recent version from PyPI.

Setup

1. Download your data via Google Takeout
2. Extract the .tgz archives into a folder
3. cd into that folder — the tool expects this layout:

your-takeout-folder/
└── Takeout/
    └── Mail/
        └── All mail Including Spam and Trash.mbox

On first run, a SQLite index is built automatically (~2 min for 8GB). After that, all searches are instant.

Install (optional)

For repeated use, install permanently so search-email is always available:

pip install google-takeout-utils
# or
uv tool install google-takeout-utils

Usage

All examples use uvx. If installed globally, replace uvx google-takeout-utils@latest with just google-takeout-utils.

Search emails

# Search by sender (case-insensitive substring match on name or email)
uvx google-takeout-utils@latest search-email --from alice

# Date range + sender, limit results
uvx google-takeout-utils@latest search-email --after 2023-01-01 --before 2023-07-01 --from john --limit 20

# Search by recipient (searches To, CC, and BCC)
uvx google-takeout-utils@latest search-email --to alice@example.com

# Search by subject
uvx google-takeout-utils@latest search-email --subject "invoice" --limit 5

# Only emails with attachments
uvx google-takeout-utils@latest search-email --has-attachment --from bank --no-body

# Count matches
uvx google-takeout-utils@latest search-email --count --from newsletter

# Full-text body search (slower — seeks into mbox for each candidate)
uvx google-takeout-utils@latest search-email --body "project proposal" --limit 5

# Headers only, no body preview
uvx google-takeout-utils@latest search-email --from alice --no-body

Search results are sorted by date (newest first). Each result shows a database ID and an [A] marker if the email has attachments.

Read a single email

# Show full email by database ID (from search results)
uvx google-takeout-utils@latest search-email --show 4521

# As JSON (useful for piping to other tools or LLMs)
uvx google-takeout-utils@latest search-email --show 4521 --output json

# As YAML
uvx google-takeout-utils@latest search-email --show 4521 --output yaml

--show displays the complete body, To/CC/BCC recipients, and lists all attachments with their extract commands.

View email threads

# Reconstruct the full thread containing email 4521
uvx google-takeout-utils@latest search-email --thread 4521

Shows an indented tree of all related emails with their database IDs, subjects (truncated to 70 chars), senders, and dates. The starting email is marked with <--.

Extract attachments

# Save first attachment of email 4521 to current directory
uvx google-takeout-utils@latest search-email --attachment 4521-1

# Save to a specific directory
uvx google-takeout-utils@latest search-email --attachment 4521-2 --output-dir /tmp

Use --show ID first to see available attachments and their index numbers.

Index management

# Force rebuild (e.g. after a new Google Takeout export)
uvx google-takeout-utils@latest search-email --re-index

How it works

On first run, the tool scans the entire mbox file and builds a SQLite index (Takeout/Mail/index.sqlite) containing date, sender, recipients (To/CC/BCC), subject, attachment flags, and threading information (Message-ID, In-Reply-To) for every email. Threads are precomputed using Union-Find on In-Reply-To chains.

After indexing, all searches query the SQLite database and return results instantly. Body text and attachments are fetched on demand by seeking to the byte offset in the mbox file.

The index is rebuilt automatically when missing (e.g. after a fresh Takeout import).

Options reference

Search filters

Display

Actions

Index

License

Apache 2.0 — see LICENSE.