Waymark · the editorial platform for scholarly journals

Every step marked.
Every number explained.

A manuscript submission and peer-review platform built from an append-only event log, not the row-overwriting model most publishing platforms still run on. The record only ever gains entries — nothing a journal has ever recorded can be silently overwritten.

Open the demo

https://waymark-demo.onrender.com/?invite=tessera-quire-46a73a

The link is the whole login. Nothing to install, nothing to create, no password to keep.

Read the five commitments
Submitted · QC'd · assigned · reviewed · decided — every waymark stays on the record.

The failure this fixes

Most publishing platforms don't lose data. They overwrite it.

Where a journal's reports don't add up, the reason usually isn't the reports — it's that the platform underneath them updates rows instead of recording history.

Assignment history

When a first round of reviewer invitations is declined and a second round goes out, the second date replaces the first. The associate editor who handled a slow first round looks one to two weeks slower than they were — the original date is simply gone.

Reproducibility

A manuscript that later moves out of the state a report excluded it from can reappear retroactively, so the same report run a month apart returns different counts. On Waymark, that same report is answerable as-of any date and returns the same numbers every time.

The message hierarchy

Five commitments, each one a specific failure made unrepresentable

These aren't aspirations. Each is falsifiable in a five-minute demo, and each names a behavior other publishing platforms still have, that this one must not reproduce.

01

History is append-only

No state transition overwrites its predecessor. A correction is a new event sitting next to the one it corrects, not a replacement for it.

the assignment-date bug above is structurally impossible on an append-only log
02

Reports are reproducible

The same query over the same range returns the same answer forever. Every report is answerable as-of a point in time, and states which point it used.

re-run any report against last month's as-of date and get last month's numbers, exactly
03

Every number carries its formula

Acceptance rate and turnaround time are the two figures everyone asks for and almost no submission system defines. The population counted and the arithmetic used are published next to the figure, not buried in documentation.

every report ships its own Method and Provenance panel
04

Defaults do not hide data

Nothing a journal would want counted is opt-in by accident. Any active filter is shown alongside the result it produced.

transferred papers are included by default, not behind a checkbox someone has to remember
05

Structured state is structured

Anything staff currently track as a colored sticky note or a spreadsheet convention becomes a typed, queryable, auditable field.

a shadow data model built from sticky notes and side spreadsheets becomes part of the record itself

One more, for the people it happens to

Nobody has to go find an email to know where things stand.

The five commitments above are about the record. This one is about the two people waiting on it — stated separately because it runs in both directions at once.

For authors

Checking in doesn't mean writing in

A submitting author sees where a manuscript stands in language that means something to them, not an internal queue name — what it's waiting on, and how long that stage usually takes. Every message about the manuscript lives in one place, so an update never requires an email asking for one.

requirements A17–A20: a plain-language status view, what's being waited on, and all correspondence in one place
For editorial staff

Chasing is a system, not a memory

Chasers to whoever owes the next action go out automatically, on a configurable schedule staff can see. Nothing is chased twice by accident, and nothing owed slips because the reminder lived in someone's head instead of the record.

requirement E27, demonstrated in Waymark Reports' Chasing tab: what's owed, who was chased, and what came back

What doesn't change

The workflow survives. The record underneath it changes.

The folder-and-queue model, consultations, and visible paper movement are worth reproducing faithfully — an editor's day should stay recognizable.

Folders and queues The same working folders — QC, awaiting assignment, under review, awaiting decision — stay the shape of the day.
Consultations An editor-in-chief can still pull in another editor's opinion on a manuscript before deciding.
Visible paper movement Where a manuscript is, and how it got there, stays a first-class view — not a report someone has to run.

The honest alternative

Why not just fix the reports?

A reporting layer bolted onto an existing platform is the closest, most credible alternative, and it can genuinely help — it can snapshot an unstable system's numbers and make a given query stable.

It has a ceiling, though: it cannot fix the reason the numbers were unstable in the first place, because that reason lives in the platform underneath, which updates rows instead of recording events. Nor can a layer touch the workflow complaints — editorial staff have said plainly that those need the system of record itself to change.

Waymark takes the other path: get the event history right first, and let reporting, audit, and reproducibility follow from a data model most publishing platforms never had.

Where this actually stands

Built, demonstrated, and still open

Running today, against synthetic data

  • The append-only event log, projection, and as-of query
  • Server-side authorization by grant, resolved per request
  • The editor surface's core loop — queues, manuscript detail, search, workflow tasks
  • A first reporting view with its Method and Provenance panel
  • Automatic, scheduled chasers and a Chasing report on what's owed and what came back

Still open

  • Author submission, reviewer invitations, decision letters, and consultations are specified, not yet built
  • Whether this ships as a production system or a public demonstrator is undecided
  • Historical migration from whatever a journal runs on today is unscoped

The running demo spans an invented family of six journals and entirely synthetic manuscripts, authors, and reviewers — built to show the model, not to stand in for any journal family's real records.