Ideas do not arrive at a desk. They arrive on a walk, halfway through someone else's sentence, or in the ninety seconds before sleep. The cost of losing them is not the idea itself — it is the version of it you had at that moment, with the connection that made it interesting. Capture has to be fast enough to happen at all; shaping can wait.
FOR CREATORS
Catch the idea now. Shape it later.
Capture ideas as voice notes, pull quotes from books and screens with OCR, and shape the raw material into a draft you can export.
JOBS TO BE DONE
What you are trying to get done
Capture the thought before it goes
Speak it. A voice note takes ten seconds and preserves the phrasing you had, which is usually the part worth keeping.
Turn spoken fragments into text you can work with
Transcription makes a voice note editable and searchable, so a month of fragments becomes a body of material rather than an audio graveyard.
Collect quotes and references
Photograph a page or a screen and OCR turns it into text in the note, with your own comment underneath it.
Get from notes to a draft
Ask AI to group what you have on a theme, or to draft an outline from the fragments you select. You edit, it does not publish.
IN PRACTICE
How this works across a session
01
Capture
One inbox, whatever the form: typed line, voice note, photo of a page. Sorting at capture time is the fastest way to stop capturing.
02
Review
Once a week, read what accumulated. Give the keepers a real title, throw the rest away. This step is where a pile becomes material.
03
Shape
Pull the relevant notes together, let AI propose a structure, then write over it. Export to Markdown, PDF, or DOCX for wherever it gets published.
Why voice is the fastest capture, and text is the only usable storage
Speaking is roughly three times faster than typing on a phone, and it works while walking. That is why voice notes get made when typed notes do not. But audio is unsearchable and unskimmable, so a folder of voice memos quietly becomes a place where ideas go to be forgotten.
Transcribing at capture time resolves the tension. You keep the speed of speaking and the usability of text, and the original audio stays available for the rare case where tone or exact wording matters.
- Record in the moment, edit at the desk
- Search the transcript by the phrase you half-remember
- Keep the audio; read the text
A weekly review is what separates a note system from a pile
Every capture tool works for a fortnight. What breaks is the review: without a regular pass that promotes some notes and deletes others, the volume grows until searching feels like archaeology.
The review does not need to be elaborate. Read what arrived, retitle anything worth keeping so it is findable by the words you will actually search for, and delete the rest without ceremony.
Where AI helps a writer, and where it does not
AI is good at the parts of writing that are sorting: grouping fragments by theme, proposing an order, telling you what you already said twice. It is poor at the part that makes writing worth reading, which is having a specific view.
The workable division is to let AI restructure material you wrote and keep the sentences yours. MemoFlow AI reads only the notes you select, and shows important changes for confirmation before they touch a note.
Worth knowing: AI output can be wrong or bland — treat it as a draft to edit, not a finished text. Verify any quote captured by OCR against the original source before publishing it.
Where to go next
Be the first to know when MemoFlow is ready.
No newsletter cadence—only the launch and essential product information.