Digital ledger · Prototype

Prototype · core workspace needs further validation

Papyr

Papyr explores how people who keep records in paper books could move to a searchable digital workspace without losing the feel of writing. The repository contains a Next.js application, book-management screens and a layered ledger canvas.

Papyr’s public sign-up page alongside a sample ledger preview, captured from the live deployment.
Live public sign-up screen, captured 3 November 2026. The pictured ledger panel is part of that screen; no account was created and the signed-in workspace was not accessed.
INTENDED USERS

People and small-business operators who prefer paper books for routine records.

This audience is the product brief’s hypothesis; the personas and success thresholds in the repository are plans, not verified customer research or outcomes.
PRODUCT STATUS

Prototype · core workspace needs further validation

TECHNOLOGY

Next.js 14ReactTypeScriptSupabase

01 / Business problem

Why this product exists.

Paper records can be difficult to search, back up and reach across devices. A digital replacement must make those benefits available without making everyday writing feel foreign.

02 / Product objective

What it is trying to change.

Organize books and pages in a digital workspace while keeping handwriting as the intended primary interaction.

03 / Solution architecture

Separate concerns.
Keep the path clear.

Module boundaries and data paths observed in source; this is not a live deployment diagram.

  1. 01

    Next.js application and authenticated routes

  2. 02

    React book and ledger workspace

  3. 03

    Separate paper, grid and ink canvas layers

  4. 04

    Supabase authentication and book data

04 / Engineering decisions

Choices worth
making explicit.

These notes describe the implementation found in the repositories—not measured product outcomes.

01

Paper, grid and ink are drawn on separate canvas layers so each visual surface can be initialized and redrawn independently.

02

Canvas setup uses the container’s measured dimensions, device-pixel-ratio scaling and a ResizeObserver; the repository’s older issue register still records a zero-dimension regression, so this path has not been certified here.

03

The ink hook builds strokes through a dedicated renderer and keeps history snapshots for undo and redo rather than treating each stroke as an unrelated image.

05 / Selected implementation

Details you can
follow to source.

The current book list fetches records for the signed-in user and provides search, recent filtering, create and delete flows. In the inspected source, the displayed page count is still set to zero as a TODO rather than queried from the pages table.

06 / Current status and limits

What remains
unfinished.

  • The public deployment exposes the entry and sign-up experience; this review did not create an account or inspect a signed-in ledger workspace.
  • The repository’s KNOWN_ISSUES.md (last updated 2026-08-16) lists a critical blank-canvas sizing regression, mobile workspace rendering issues, and handwriting recognition and recognized-text persistence as unfinished. These dated issues were not reproduced or re-tested in this task.
  • The MVP document’s latency, adoption and usability figures are targets, not measured results. No user adoption or performance outcome is claimed here.

07 / Potential improvements

What to do
next.

  • Reproduce and close the canvas-sizing and mobile-layout issues on current devices.
  • Complete and test recognition-to-cell persistence before presenting recognition as a feature.
  • Add an authenticated, non-production demo with safe sample records and a documented reset path.
Back to all engineering work

A useful next conversation

Have a problem
like this?

Bring the situation, not a finished technical brief.

Discuss a project