How to export your Day One journal (step by step)
Day One exports in three formats and only two of them are useful outside Day One. Here is which to pick, how to get it, and what survives the move.
Getting paper journals into a searchable form, without spending a month on it. What to photograph, what to type, and what to leave on the shelf.
People with paper journals tend to fall into two groups. One group has no intention of digitizing and is entirely happy. The other has been meaning to do it for years, feels mildly guilty about it, and has not started because the job looks enormous.
If you are in the second group, the job is smaller than you think — but only if you are willing to not do all of it.
The right method depends completely on the goal, and most stalled projects stall because the goal was never named.
To preserve it. You want the words to survive a flood, a move, or a fire. This is a photography problem. You do not need text; you need images, backed up.
To search it. You want to find the entry about the thing. This requires text, which means transcription of some kind.
To read across it. You want to see what recurred over fifteen years. This requires text and dates, and it is the most demanding version.
To make something from it. A letter, an essay, a book. You need text for the parts you will use and nothing at all for the rest.
Only the third goal requires doing the whole archive. The others are satisfied by doing part of it, which is why naming the goal first saves so much time.
Whatever your goal, photograph everything before you transcribe anything.
Photographing is fast — a notebook takes a few minutes — and it is the step that protects you against loss. Transcription is slow and can happen over years, on whatever subset turns out to matter. Doing it in this order means the irreplaceable part is safe within an afternoon.
Practical notes, learned the boring way:
Then back the images up in two places before you do anything else.
This is the part people expect to be solved and are surprised to find is not, entirely.
Machine transcription of handwriting has improved enormously and is now genuinely useful. It is not perfect, and it is least reliable exactly where you most want accuracy: hurried writing, idiosyncratic shorthand, names, and the cramped note squeezed into a margin at 1am. The Smithsonian still runs a volunteer transcription programme for handwritten historical documents, which tells you something about where the hard edge remains.
What this means in practice is that you should treat machine transcription as a first draft that is usually right, not as a finished text. For a personal journal that is generally fine. You are not producing a scholarly edition; you are producing something searchable.
Ember has this built in — you can photograph handwritten pages, up to ten at a time, and it transcribes them into entries you review before anything is saved. The review step is the important part of that sentence. Anything that writes directly into your archive without showing you the result first is doing you a disservice, because you will never find the errors afterwards.
If you do one thing carefully, do this.
A transcribed archive without dates is a pile of text. A transcribed archive with dates is a timeline, and almost everything useful you might later do with it — reading by period, following a thread, seeing what changed — depends on knowing when each entry was written.
Paper journals are inconsistent about this. Many entries say "Tuesday" or nothing at all. Where the date is missing, infer what you can from surroundings and record your uncertainty rather than guessing silently. "Approximately March 2014" is far more useful later than a confident wrong date.
When you bring text into an app, check how it handles dates. Ember splits a pasted block into separate entries at each date it recognises, and it reads a range of formats — "July 15, 2025", "15/07/25", "2025-07-15", or bare "July 15" — which means a transcript that keeps its date headings converts cleanly into a real archive rather than one enormous entry.
Here is the approach that most often actually gets finished.
This is unsatisfying advice if you wanted permission to spend a month on a complete archive. But complete-archive projects have a poor completion rate, and a half-finished one abandoned in month two is worth less than one notebook fully done.
The reason to do any of this is that paper and text behave completely differently once you have fifteen years of them.
Paper is excellent for one notebook. You can hold it, flip it, find the entry by the shape of the page. Paper is useless across fifteen notebooks, because there is no way to ask a question of a shelf. You cannot find every entry that mentioned your father. You cannot see that one subject ran continuously from 2011 to 2019.
Text can be asked questions. That is the entire difference, and whether it is worth the work depends on whether you have questions. If you do, reading across a whole archive is what the transcription was for — and it is worth knowing that the value arrives only once there is enough text for a pattern to show up in.
If you are in the middle of a chapter you expect to want back later, there is also a case for writing the next years differently rather than only rescuing the old ones.
Do not throw away the notebooks.
Transcription loses the marginalia, the pressure of the pen on a bad day, the page you tore out and the stub it left. None of that is searchable and some of it is the most direct record you have. Digitize for use; keep the paper for the things text cannot hold.
Day One exports in three formats and only two of them are useful outside Day One. Here is which to pick, how to get it, and what survives the move.
A long archive is either an asset or a folder you never open, and the difference is what you do next. Four honest options for a decade of writing.