GxPlex

How this is kept

This page answers two questions without sending you anywhere else: can this record be trusted, and how would you check it yourself.

Evidence and interpretation are separate layers; this record publishes the first, and keeps the second on the reader’s side.

The record publishes what official sources published, when, how it relates, and what is forming — and never what should be done about it.

What is tracked

132 regulations, guidance documents and standards from 31 publishers across 10 jurisdictions. Each one is fetched, read and dated every day. The series begins 2026-07-28 and has run for 53 days.

For each document this record holds what the publisher's own page said, the day it was read, the address it answered at, and — where the text is public — a digest of that text so a later change can be proved rather than asserted.

What is not tracked, and what is not yet determined

“not yet determined” is a published value here, not a gap. The schema requires a status and this record writes it deliberately rather than guessing. These entries reach it six different ways, and they are not the same claim:

0 machine located — none today
A tool located a passage on the publisher’s surface and proposed no value — the words that locate it are publisher vocabulary this enum does not receive without a reading.
2 surface states none machine
A tool read the surface and it states no status term this publisher’s family declares. The read is recorded with its event; no person has checked it.
1 surface states none person
A person enumerated the publisher’s surfaces and found no status stated. Bounded to the surfaces named.
6 withdrawn attestation
This archive published a value here and took it back, because nothing attested it. The withdrawn value is kept on the entry with its date.
40 unmeasured with cause
No request was made — the host’s request budget was spent before this entry was reached. Nothing was measured, and that is different from finding nothing.
25 not yet read
Nobody has looked yet. No extraction has run for this entry’s status.

— 74 entries, which is what these six decompose.

Some documents cannot be read in full, and the reason is recorded on each one.

A further 44 are not re-checked, and the reason each one records is: 23 paywalled — progression tracked; 9 listing only; 6 not a document; 3 blocks automated requests; 3 register states no status. Ordered by count, then alphabetically.

A document that cannot be read is still listed, with the reason, rather than quietly dropped. An absence you can see is worth more than a tidy list.

Why a status is not yet determined

74 of 132 entries carry no determined status. That is this record’s largest single fact about itself, and every one of them records why:

Why these 74 are not yet determined. Ordered by whose share each names: this record’s own first, then the publisher’s, then the rows whose records name no share at all.

  • 0 — this record has not reached them
  • 26 — this record promises no person-verification for them
  • 48 — the publisher does not say, or will not permit access
  • 0 — nothing held names a reason. This is an absence of a reason, not a kind of one

What these reasons are recorded against

Four of the reasons below are recorded against what a publisher has issued and what it permits, rather than against how far this record has read: published only within a listing, no public document, the site refuses automated requests, and a register that carries no status. As recorded, 21 of the 74 entries shown here carry one of the four in their own recorded reason field; a further 0 of the archive’s 132 entries carry one and are not shown here. 1 of the 21 also carries checked_by: human with a check date, and still reads unknown.

Counted by the classes listed below rather than by the field, the figure is 18. Three of those classes take every member from the field. The fourth value shares a class with a filing label: that class holds 26 entries, 3 carrying a register reason and 23 carrying record_tier: register and no frozen reason at all. The two populations are disjoint, and no count here separates them.

The frozen reason paywalled_stage_tracked is in neither figure: 23 entries carry it, and 10 of those hold a status other than unknown.

Both figures are read from fields, and each moves when its field moves — after a publisher publishes a document this record has recorded as unpublished, issues one outside the listing that currently carries it, or changes what it serves to an identified client, and whenever the reason itself is re-derived from a later reading.

Every reason below is read from a row’s own recorded field or from an outcome its retrieval ladder recorded — never from this archive’s own prose about the row. No status is inferred: a row with nothing recorded is counted as having no recorded reason.

not yet requested — 0
the ladder made no request — this run's budget or a robots policy
register entry — no confirmation promised — 26
record_tier: register, or frozen_reason: register_no_status
text behind a paywall — 13
frozen_reason: paywalled_stage_tracked
published only within a listing — 9
frozen_reason: listing_only
no public document — 6
frozen_reason: no_public_document
access blocked to an identified client — 3
frozen_reason: access_blocked
read, and it states no status — 17
a rung at a publisher-published address fetched the document; no status passage was found
the address returns not-found — 0
a rung at a publisher-published address received HTTP 404 or 410
the request was refused — 0
a rung at a publisher-published address received HTTP 401, 402 or 403
no reason recorded — 0
nothing held names a reason — this is an absence of a reason, not a kind

What the verification marks mean

The board carries a verification column. Each mark is defined here — and it is defined SOMEWHERE, on purpose: the same labels are drawn onto the social cards, which are images and cannot carry a correction. A label that appears only on an uncorrectable image is a claim with no place to be fixed.

These count what this record has done about each entry, not what the document says about itself. The document's own status — draft, final, in force, withdrawn or unknown — is a different axis over the same entries, published per row and in the API.

How the two axes overlap
Every tracked entry appears once. Rows: what this record has done about the entry. Columns: the status the document states about itself. Both margins count the same 132 entries.
verification \ status unknownfinalin-forcedraftwithdrawnsuperseded all
verified 03911320 55
frozen 3400000 34
register 2302000 25
awaiting 1600000 16
tool 000001 1
absence 100000 1
all 743913321 132

No cell is marked good, bad or notable. The grid states how the two counts overlap and nothing about whether any cell should be larger or smaller.

What these states mean
verified
checked against the primary source by a person
tool-read
read from the source by tooling, because the publisher prints no status a machine can read
checked — no status stated
a person checked the publisher's own surfaces and none states a status
awaiting
not yet checked
not re-checked
no status recorded; each row states its reason
a register
observed daily; no person-verification is promised
Every verification state, including those with no entries today
verified
tool-read
awaiting
checked — no status stated
check unrecorded
not re-checked
a register

Nothing here is interpreted

This record reports what a document says and when it changed. It does not say what a document requires of you, whether it applies to your product, or what action it calls for. There is no advice here, no assessment, and no ranking.

That is a deliberate limit, not an unfinished feature. A reported status is checkable against the publisher's own page; an interpretation is not, and a record that published both would make the checkable half harder to trust.

Who keeps it

One person, working alone, with the collection automated and every published status written by hand. No organisation stands behind it and no employer is involved.

What survives that does not depend on them: the underlying record is plain text, one object per line, with the digests and the dates in the open. It can be read, checked and rebuilt without this site and without its code.

What this site stores on your device. Two controls on the board — compact rows, and your own list of regulations — save their setting in your browser’s local storage under keys beginning gxplex-, only when you use one of them; nothing is stored on a visit that touches neither, there are no cookies and no accounts, nothing saved is ever sent anywhere or read by this site’s servers — every visitor is served identical bytes — and Forget this device on the board clears all of it in one tap.

When something here is wrong

The corrections policy. A published claim about a document — its status, a date, a quoted span, a recorded value — is never silently edited: the correction is added with its date and the wording it replaced, and it stays on corrections permanently; corrections to this record’s own machinery, such as a counting rule or a page template, are kept in the repository history instead, because they change nothing this site has said about any document.

Factual claims are never edited silently. A correction is added, dated, and kept beside what it corrects — 0 correction(s) recorded so far. A record whose past statements can change without trace cannot be cited, because a reader who quoted it last month has no way to know whether the quotation still stands.

What gets corrected: any published factual claim — a status, a date, a digest, a count, or a sentence that overstated what the record establishes. How: the original stays, the correction is added beside it with its date and what it changed. Where: the full log, and the feed that carries it, are on Developers.

An error can be sent to corrections@gxplex.com. There is no mechanism that would record whether a correction originated outside this record, so no count of that kind is published.


What gets included, in full

Without a stated test a corpus grows by association until the intersection stops being an intersection. This is the test.

A document is tracked if it governs, in a GxP context, one or more of:

A document is not tracked merely because it is cited in that context. ICH Q7, Q8, Q10 and Q11 are quality-management documents. Tracked documents reference them constantly; they do not themselves govern computerised systems, and being frequently cited is not the test.

Applying the rule is a judgement, and this record does not claim to be complete against it. The daily run observes the entries it already holds; it does not search for documents it does not have. An entry is added when a person applies the test above and records the reason. So a document that satisfies one of the four limbs may be absent from this record for either of two reasons, and they are not the same: it was considered and excluded, in which case the exclusion is stated, or it has not been considered. An absence here is not evidence that a document does not exist, is not in force, or does not govern.

The most recently added entry began being observed on . That date moves when an entry is added and not otherwise, so a corpus size that has not changed is a record of decisions not taken rather than of a process that stopped.

ICH Q9(R1), Q12, Q13 and Q14 are tracked, because quality risk management, lifecycle change management, continuous manufacturing and analytical procedure lifecycle each bear directly on computerised control and on the data that control produces.

Documents and watch entries

The rule tests documents. The record also carries watch entries — a committee, a work programme, a guidance agenda, a register of authorised devices. These do not govern anything and are not measured against the rule; they are tracked because the documents they will produce fall inside it. Every one is frozen with a stated reason, so the two are never counted as one.

Software as a medical device

In scope. Software as a medical device is governed by the same computerised-system and lifecycle questions as manufacturing software, and in several jurisdictions by the same authority. EU MDR, EU IVDR, the IMDRF SaMD documents and the FDA device guidances are tracked on that basis, and the device-side standards are tracked with them.

Admissions are drawn from EU, US, UK and international publishers. Where a document from another jurisdiction is held, it predates that boundary and is kept rather than removed — dropping it would be a silent edit to what this record has already published.

The licence boundary

The text of a paywalled standard is not taken. Not a sentence, not a normalised digest of the body, not a diff. What is followed instead is the publisher's own catalogue entry, which is public and which states the stage.

This costs the record something real: for those 23 documents a change in the text itself cannot be detected, only a change in what the publisher says about it. That cost is stated rather than worked around.

The five positions this record holds
  1. The text of a paywalled standard is not taken. The licence is a boundary, not an obstacle to work around.
  2. A digest records that text changed, not that meaning changed. A typo and a substantive rewrite produce the same signal, and this record does not tell them apart.
  3. Search matches the words this record published, and nothing else. There is no synonym expansion and no ontology mapping one body's vocabulary onto another's.
  4. A status is written by a person pressing a key. Collection is automated; judgment is not, and no machine writes a status here.
  5. It is kept by one person, and what survives that does not depend on them. The format is open and the record outlives its keeper.
Checking a line yourself

Every regulation page carries the address the document was read at and the day it was read. Open the publisher's page and compare it with what is quoted here. That is the whole check, and it needs nothing from this site.

Where the text is public, a digest of it is recorded for each day. Two days with the same digest saw the same text; two days with different digests did not. The digests, dates and addresses are all available as data — see Developers.

What happens to this record if it stops being kept

Stated as intent, not as a guarantee. The record is plain text in an open format, and it is intended to remain readable and rebuildable by anyone who has a copy, without this site and without its code. Nothing about that is contractual, and no arrangement is in place that would compel it.

The board · What changed · Developers

Last updated: