Importing your archives

Much of a person’s history has no live API. It sits in export archives — a Google Takeout, an Apple privacy download, a Twitter archive, a folder of photos off an old drive. Importing is how that history gets into your Companion Memory.

You upload an archive, or copy it into the inbox folder. Companion Memory expands it, reads it with a parser that understands that particular export, and saves what it finds as messages, events, content, things you own, places you have been, and browsing history.

What the catalogue covers

As of Companion Memory 2026.10.2 (3 October 2026), the Import page carries an export guide for 105 sources. Of those, 85 have a working parser. The other 20 carry a Coming soon badge: the guide still tells you how to request the export, and an archive you keep imports without a new export once the parser ships.

Some services share a parser. The guides for OneDrive, Dropbox, Backblaze, Koofr, Synology and a local media folder all use the parser for loose photos, videos and documents. Behind the 85 ready guides sit 77 archive parsers, including that catch-all. One more importer, BlueBubbles, reads a live iMessage bridge instead of an archive — you set it up under Sources → Services.

These numbers change as parsers ship. The Import page always shows the current catalogue.

Everything happens on your hardware

The archive is stored on your computer, expanded there, and read there. Nothing about its contents is sent anywhere else while the import runs.

Two things can reach outside afterwards. Some exports reference media by link rather than including it — Reddit’s export does. Companion Memory saves the link and not the file, and when you view that record your browser loads the image or video from the service that hosts it. And once your records are saved, background jobs can look up places and link previews from outside sources — see What leaves your appliance.

What you get out of it

Imported data is parsed to generate Memory’s core data types:

• Messages — emails, texts, chats, calls, with the people who sent and received them attached.
• Content — media such as photos and videos, as well as documents worth searching: notes, bookmarks, articles, files.
• Events — things that happened at a time: a calendar entry, a purchase, a ride, visiting a location.
• Things — physical items you bought, in your inventory.
• Locations — GPS fixes, whether from a location history or imprinted on a photograph.
• Browsing — pages you visited, from browser history exports.
• People — everyone you interacted with in the records above.

One record from an import can produce several of these Memory records. A geotagged message is both a message and a location fix. A photograph with a date is both a moment on your timeline and content you can search for.

What it is not

Not a backup. Keep your original exports. This is an archival and organisation tool, and while it is under active development you may need to import an archive again to get the benefit of a parser improvement.

Not encrypted at rest by itself. Your imported records live in Memory’s database and object storage like everything else it holds. Use disk encryption if you need protection against someone accessing your device.

Where to go next

• Prepare your data — start here, because some services take a week to hand over an export.
• Importing data — the walkthrough, for both the wizard and the inbox folder.
• How importing works — what the four stages do, and what runs after the import ends.
• Import FAQ — stalled bars, partial results and re-importing.