We opened every window before breakfast. The place still smelled like cedar, dust, and the lemons growing by the wall.

By afternoon, the garden was louder than the house again. I want to remember how ordinary—and how complete—it felt.
A place for a life in progress
Your personal archive for the things you live, learn, make, and remember—kept private, shared with a few, or opened to the world.
Your life is more than a feed
Magpie is a calm place to collect the moments, ideas, places, people, and pursuits that make a life yours.
Not a performance. Not another endless scroll. A living archive you can return to—and let others into, on your terms.
As the internet fills with machine-made content and feeds train us to skim, scroll, and perform, Magpie makes room for slower human acts: reading, writing, remembering, and thinking in your own voice.
An editor that moves with you
Spill out a few words or follow a thought wherever it goes. Upload, resize, and position images; drag blocks beside one another or stack them vertically—all without breaking your concentration.
This is the real editor—try it. Drag a block, type “/” for commands, drop in an image.
Privacy at every scale
Privacy is not one switch at the top of your profile. Keep a whole section to yourself, limit one entry to family, or hide a single paragraph inside an otherwise public story.
A whole sectionKeep an entire collection—like family or work—private.
One entryChoose public, unlisted, followers, selected people, or only you.
Any blockHide a photograph, paragraph, or detail without hiding the story.
Sunday at the old house
Only meWe opened every window before breakfast. The place still smelled like cedar, dust, and the lemons growing by the wall.

By afternoon, the garden was louder than the house again. I want to remember how ordinary—and how complete—it felt.
Neural networks, from signals to meaning
Public entryA neural network learns by adjusting the strength of the connections between layers—not by storing a ready-made answer.
# a small forward pass
hidden = relu(x @ weights_1 + bias_1)
prediction = softmax(hidden @ weights_2)Training repeats this process, measuring error and sending corrections backward through the network.
Your life, across time
Milestones, photographs, notes, and knowledge become small markers along your timeline. Drag through the years; choose one to let the moment expand again.
Learning the shape of an early wave
Drag the timeline. Select a moment to open it.
Your archive

“Everything became quiet near the edge…”

“The sea was almost empty…”
“I keep returning to the same idea…”
Begin anywhere