Email Collection Toolkit

Email Collection Toolkit

Archive Gmail

Import a Gmail snapshot using Google Takeout and MBOX.

END USER

Use Google Takeout today. This does not grant Email Collection Toolkit access to Gmail and does not require an OAuth client.

  1. Choose Deselect all, then select Mail.
  2. Retain all mail data for the strongest completeness evidence.
  3. Choose a one-time ZIP export and Create export.
  4. Download every part when Google emails you, then extract the files beneath one directory.
  5. Import the extracted directory into Email Collection Toolkit.

Google exports message content, headers, attachments, and labels. Takeout is a snapshot rather than an incremental service, and large exports may be divided into several ZIP files. Keep the ZIP files until the resulting archive passes verification.

For best-effort updates between Takeout snapshots, an account already synchronized with Apple Mail can be ingested through its local cache. This can be rerun as more complete messages arrive, but it is not proof of server completeness. See Advanced for cache limits, exact deduplication, and semantic-message reconciliation.

DEVELOPER

The canonical Gmail design and operations document covers Takeout ingestion, the future Gmail API adapter, shared Desktop OAuth clients, Testing versus production, personal-use exceptions, restricted-scope verification, annual assessments when server-side data handling makes one necessary, and why IMAP is not an authentication shortcut.

The illustrated one-time registration reference remains available for maintainers experimenting with the future API adapter.

Current acquisition boundaries

No release Desktop OAuth client is bundled yet. The shared-client end-user flow remains deferred until a maintainer supplies and validates that public configuration in release artifacts. Current authorization requires a developer client override. Installing that override validates and writes the same bytes. Known consumer domains need no DNS lookup; transient DNS and token-refresh transport failures are disclosed as errors rather than negative detection or fresh consent. Credentials are stored only after the profile matches.

Directory import reports and skips .partial.emlx records and imports complete supported messages. Direct selection of a partial record is rejected. Detached content is not reconstructed, so this remains best-effort cache recovery rather than complete provider acquisition.