AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 18 November 2026

Does Landing AI keep footnotes out of the body flow and attach them to their references?

Landing AI kept footnotes outside body paragraphs while preserving their references. It retained all thirteen notes intact, grouping each page’s notes after its body text.

1 of 1 test case passed

Every test case in this scenario has a result.

Pass rate100%1 of 1 gradable
Coverage1 of 1test case with a result
1PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the test case below: what we sent, what we checked, what the tool returned, and the proof.

The test case

Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.

Footnote — out of body flow, associatedPassEvidence
What we sent

We submitted the complete four-page journal offprint to Landing AI’s document-to-Markdown conversion API in one request, without model or options overrides.

InputDocument (PDF)
Input PDF: journal offprint with bottom-of-page footnotes and inline references
Open the PDF ↗
What the tool returned

Copied from Output file (Markdown), lines 56–62

Markdown
<sup>1</sup> Braithe Estuary Order 1974, art. 6(2). The Order is reproduced in full in the appendix to R. Sandiman, The Braithe Estuary:
A Documentary History (Vardenne, 1988), 211-24.
<sup>2</sup> Minutes of the Harbour Board, 14 May 1973 (Braithe Record Office, BHB/3/17), 4.
<sup>3</sup> The point was not argued in Tarbert v Braithe Harbour Board [1981] 2 EstLR 88, where both parties proceeded on the
assumption that art. 6(2) applied.
<sup>4</sup> The Board's file on the 2011 consents (Braithe Record Office, BHB/11/4) contains three letters returned undelivered and a
note of a telephone call to a solicitor in Winnipeg.

Copied from Output file (Markdown), lines 130–207

Markdown
5 Report of the Committee on Estuarine Consents (Cmnd 6712, 1977), paras 88-94.
6 See further A. Prosser, 'Consent and its conditions' (1994) 12 EPR 3, 9-14, whose account of the 1977 committee differs
from mine at one point of substance.
7 The Harbour Board series appears to have been reconstructed in 1981, probably for the purposes of the Tarbert litigation.


<!-- PAGE BREAK -->

THE 1974 BRAITHE ORDER AND THE LIMITS OF CONSENT


145


The committee's papers are in two series and neither is complete. What survives of the drafting
is a bundle of eleven documents: four memoranda, the minutes of six meetings, and a marked-up
print of the draft Order with two hands on it, one of them the secretary's.<sup>8</sup>


The sequence that matters begins at the second meeting, on 14 May 1973, at which the
solicitor's memorandum was considered and not adopted. The minute is short and gives no reason,
which has led more than one writer to treat the decision as a formality.<sup>9</sup> The marked-up print
shows that it was not. The words 'the owner for the time being' are struck through in the
secretary's hand, and beside them, in the other hand, is written 'not without the tenancy point'.


That note is the hinge of this article. It ties the deletion to the tenancy question, which the
committee took up at its fourth meeting and settled in a way that makes the wider reading of
article 6(2) impossible. The connection appears not to have been noticed before, and the tenancy
discussion is worth setting out at length.<sup>10</sup>


The problem the committee faced was this. Much of the foreshore between the training wall
and Thrimby Point was let, in 1973 as now, on grazing tenancies of a year and less. If consent ran
with the land, a consent given by an owner would bind a tenant who had not been asked; if it did
not, the Board would have to seek consent from whoever held the land at the date of the works,
tenant or owner. The committee wanted neither result and said so in terms.


8 Braithe Record Office, BHB/3/14-24. The marked-up print is BHB/3/22.
9 Notably Prosser, 'Consent and its conditions', 7, who calls the deletion 'tidying'.
10 Minutes of the fourth meeting, 9 July 1973 (Braithe Record Office, BHB/3/19), 2-6.


<!-- PAGE BREAK -->

146


HELENA M. CROYDE


Its solution was article 6(3), which nobody reads: consent is required from 'the owner and,
where the land is let for a term of more than one year, the tenant'. The threshold of a year is the
committee's own and appears nowhere else in the Order.<sup>11</sup>


Article 6(3) only makes sense on the narrow reading. If consent ran with the land, a tenant's
consent would run with the tenancy, and a threshold of a year would protect nobody: a grazing
tenant of eleven months would be bound by a consent given by a predecessor's landlord, which is
precisely the result the committee said it wished to avoid. On the narrow reading the threshold
does what the minute says it was for.


There is a further point, smaller but harder to answer. The committee considered and rejected a
draft in which consent was expressed to bind 'successors in title'. The rejected words are in the
same marked-up print, struck through in the same hand.<sup>12</sup> A committee that struck out 'the owner
for the time being' and 'successors in title' at consecutive meetings was not tidying.


The remainder of the drafting history adds little. The Order was settled at the sixth meeting and
went to the department unchanged in this respect. The department's own file, which I have not
seen, is said to contain nothing on article 6 at all.<sup>13</sup>


11 Braithe Estuary Order 1974, art. 6(3). The threshold is discussed at the fourth meeting, ibid., 5.
12 Braithe Record Office, BHB/3/22, fo. 4.
13 I am grateful to the departmental records officer for a search, and for the confirmation that the file was weeded in 1998.

Copied from Output file (JSON), metadata

JSON
{
  "job_id": "parse-01m22n0gk374he0ezb30eyazw1",
  "model_version": "dpt-3-pro-20260710",
  "page_count": 4,
  "output_markdown_chars": 11053,
  "range_units": "unicode_codepoints",
  "openapi_spec": "https://api.ade.landing.ai/openapi.json",
  "failed_pages": [],
  "duration_ms": 21403,
  "billing": {
    "service_tier": "priority",
    "total_credits": 9.6
  }
}

Returned Markdown: separate footnote blocks and inline reference markers

Open the Markdown file ↗

Raw API response: converted text and processing details

Open the whole JSON file ↗
Expected vs. Found
✓Expected: Keep all thirteen footnotes intact, outside body paragraphs and associated with the sentences that reference them.Found: All thirteen notes remain intact in separate blocks after each page’s body; thirteen inline reference markers remain attached to their sentences.
Supporting proof
Proof 1Screenshot
Comparison image: page 1 footnotes 1–4 beside the returned Markdown
Open original ↗
Proof 2Screenshot
Comparison image: page 2 source footnotes and returned inline references
Open original ↗
Proof 3Screenshot
Comparison image: page 3 footnotes 8–10 and their inline references
Open original ↗
Proof 4Screenshot
Comparison image: page 4 footnotes 11–13 and their inline references
Open original ↗
Why this result

Separate note blocks and sentence-level references meet the required structure. Note 4 is present through the end of its sentence, so its text is complete rather than truncated.

Also observed in this test case

Inconsistent note numbering style. Notes 1–4 use superscript numbers at the start of the note; notes 5–13 use plain leading numbers.

Running heads remain. The returned Markdown keeps the running heads and printed page numbers on their own lines.

Tested by Chandresh Bisht · evidence dated 9 September 2026

Configuration and setup

How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.

Software that produced the output
Landing AI
Build or version
ADE Parse API v2; model_version dpt-3-pro-20260710
Plan or tier
Explore (free tier)
Surface
API
Set up before the run
The journal offprint about the Braithe Order had to be converted in full to Markdown, with bottom-of-page footnotes referenced in the body, using the default pipeline in one pass without tool-specific tuning.
Conversion
default pipeline, single pass, no model or options overrides
Tested
9 September 2026 · Chandresh Bisht

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.

The rules
  • Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
  • Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
  • Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.

Where this sits in the benchmark

This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.

History of this result

What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.

from the publication record
19 November 2026First publishedConverting a complex PDF into clean Markdown with a hosted API v1

Act on this result

Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/pdf-to-markdown-apis/results/landing-ai/footnotes-at-the-bottom-of-the-page · 1 pass · 0 fail · coverage 1/1 · graded 2026-11-18

The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.

Verify the proof files

These files support this result. Open a file to inspect the original evidence.

Input · Document (PDF) ↗application/pdf · 152 KB
Output file (Markdown) ↗text/markdown · 11 KB
Proof 1 · Screenshot ↗image/png · 912 KB
Proof 2 · Screenshot ↗image/png · 284 KB
Proof 3 · Screenshot ↗image/png · 174 KB
Proof 4 · Screenshot ↗image/png · 150 KB
Output file (JSON) ↗application/json · 74 KB
File fingerprints (SHA-256)

A fingerprint identifies the exact file used for this result.

Input · Document (PDF)fce3c87776b6f1b8f34dcfb0f6a274b5954edcb3d33ff7ee84840dfe2afbc1ad
Output file (Markdown)55a9f551ef16ae44c45ba0026751a037ed68d6dde0afbc1bdab3dd5fc141105d
Proof 1 · Screenshotea71ae5d779d9216ac3040e420d1c0642ce123d60134d92f3f8c3c7c656bc803
Proof 2 · Screenshotaa7554248e63b1de185961fc856e0070f7020696c124aef4eb6e1e276973947c
Proof 3 · Screenshot4fa2f444a133799cd5b0386794b47751fb4b6c13bfd0278df5682df2daacb1d1
Proof 4 · Screenshot8a177d37234d2ef1e44e28518e4bf54dcb3c6abd17e89ed2a162329f876f8f4f
Output file (JSON)5cd2776977ad8e1735508e2a975b4ab5f0de1ad7f6e48b8ef2943f4191000e54