OPEN
INSTRUMENTS

SEO ARCHITECTURE · EXPLORING AN ARCHIVE

Internal linking.
Which paths lead
to your key pages?

Follow a path. Verify each link. Test a hypothesis.

Documented relations, not a ranking prediction. No crawl runs during your visit.

ARCHIVED GRAPH · 2026-09-09
3minimum links37shortest paths

Agence web, SEO & visibilité IAArchitecture SEO : comment structurer un site pour Google et les moteurs IA

View a path and its alternatives ↓

The archived example is readable below. Loading interactive controls…

01 / EXPLORE

From one point to another.

Graph archived on 2026-09-09
Package 1.1.0 · Demo 1.0.0

Choose calculation URLs

Search URL or label. The selection remains visible even when it does not match the filter. Groups distinguish documented sources from destination-only URLs.

Calculation from the root published in the manifest.

Shortest path in the archived graph

3minimum links
37shortest paths

In this archive, this destination can be reached from this starting URL by following at least 3 links. 37 shortest paths are available.

Length describes this graph only: no SEO score, indexation, traffic or visit probability. No universal three-click rule.

One path, step by step.

Deterministic order; display limited to 50 paths. The total remains exact.

  1. STARTAgence web, SEO & visibilité IAArchived H1https://www.edikka.com/Inspect this URL
  2. STEP 1accessibiliteURL-derived labelhttps://www.edikka.com/accessibiliteArchived relation · inspect evidence
  3. STEP 2seoURL-derived labelhttps://www.edikka.com/faq/seoArchived relation · inspect evidence
  4. STEP 3Architecture SEO : comment structurer un site pour Google et les moteurs IAArchived H1https://www.edikka.com/insights/seo/architecture-seo-siteArchived relation · inspect evidence

This ordered list also forms the diagram: each step follows a directed relation to the next. Full URLs remain visible.

02 / SIMULATE

What if this link changed?

LOCAL COPY OF THE GRAPH

A reversible hypothesis, without changing the site or archive. Up to eight operations; editable sources are limited to documented adjacencies.

Edit the scenario

Two hypotheses to try

These examples replace the scenario with home → French guide.

    Removing a relation deletes the entire source–destination pair from the model, not necessarily one HTML occurrence. A calculated shortcut is not a recommendation: relevance depends on the content and the reader’s needs.

    03 / EXAMINE

    Evidence, link by link.

    Relations in the displayed path. Opening a current URL does not reproduce the 9 September observation.

    Starting URL: metadata and scope

    Agence web, SEO & visibilité IA

    https://www.edikka.com/

    Source: documented adjacency.

    H1
    Agence web, SEO & visibilité IA
    Archived language
    fr
    Canonical
    https://www.edikka.com/
    HTTP / robots
    200 / index, follow
    pages.csv · line 2

    Open the current URL (new tab / nouvel onglet)

    01 Archived relation · accessibilite

    ARCHIVE · 9 SEPTEMBER 2026

    Source
    https://www.edikka.com/
    Destination
    https://www.edikka.com/accessibilite
    Reference
    graph.json → adjacency["https://www.edikka.com/"][1]
    {
  "source": "https://www.edikka.com/",
  "destination": "https://www.edikka.com/accessibilite",
  "observed_at": "2026-09-09"
}

    Anchor text: unavailable. Page area: unknown. No menu/content/footer classification is inferred.

    No additional row in the links.csv excerpt; relation evidence is in graph.json.

    Metadata for both URLs

    Agence web, SEO & visibilité IA

    https://www.edikka.com/

    Source: documented adjacency.

    H1
    Agence web, SEO & visibilité IA
    Archived language
    fr
    Canonical
    https://www.edikka.com/
    HTTP / robots
    200 / index, follow
    pages.csv · line 2

    Open the current URL (new tab / nouvel onglet)

    accessibilite

    https://www.edikka.com/accessibilite

    Source: documented adjacency.

    URL-derived label; H1, title, HTTP status, canonical and robots are not documented for this URL in pages.csv.

    Open the current URL (new tab / nouvel onglet)

    02 Archived relation · seo

    ARCHIVE · 9 SEPTEMBER 2026

    Source
    https://www.edikka.com/accessibilite
    Destination
    https://www.edikka.com/faq/seo
    Reference
    graph.json → adjacency["https://www.edikka.com/accessibilite"][20]
    {
  "source": "https://www.edikka.com/accessibilite",
  "destination": "https://www.edikka.com/faq/seo",
  "observed_at": "2026-09-09"
}

    Anchor text: unavailable. Page area: unknown. No menu/content/footer classification is inferred.

    No additional row in the links.csv excerpt; relation evidence is in graph.json.

    Metadata for both URLs

    accessibilite

    https://www.edikka.com/accessibilite

    Source: documented adjacency.

    URL-derived label; H1, title, HTTP status, canonical and robots are not documented for this URL in pages.csv.

    Open the current URL (new tab / nouvel onglet)

    seo

    https://www.edikka.com/faq/seo

    Source: documented adjacency.

    URL-derived label; H1, title, HTTP status, canonical and robots are not documented for this URL in pages.csv.

    Open the current URL (new tab / nouvel onglet)

    03 Archived relation · Architecture SEO : comment structurer un site pour Google et les moteurs IA

    ARCHIVE · 9 SEPTEMBER 2026

    Source
    https://www.edikka.com/faq/seo
    Destination
    https://www.edikka.com/insights/seo/architecture-seo-site
    Reference
    graph.json → adjacency["https://www.edikka.com/faq/seo"][24]
    {
  "source": "https://www.edikka.com/faq/seo",
  "destination": "https://www.edikka.com/insights/seo/architecture-seo-site",
  "observed_at": "2026-09-09"
}

    Anchor text: unavailable. Page area: unknown. No menu/content/footer classification is inferred.

    Additional documentary extract : links.csv · line 3 ; 1 recorded occurrence(s). The relation is deduplicated in the calculation.

    Metadata for both URLs

    seo

    https://www.edikka.com/faq/seo

    Source: documented adjacency.

    URL-derived label; H1, title, HTTP status, canonical and robots are not documented for this URL in pages.csv.

    Open the current URL (new tab / nouvel onglet)

    Architecture SEO : comment structurer un site pour Google et les moteurs IA

    https://www.edikka.com/insights/seo/architecture-seo-site

    Source: documented adjacency.

    H1
    Architecture SEO : comment structurer un site pour Google et les moteurs IA
    Archived language
    fr
    Canonical
    https://www.edikka.com/insights/seo/architecture-seo-site
    HTTP / robots
    200 / index, follow, max-image-preview:large, max-snippet:-1, max-video-preview:-1
    pages.csv · line 5

    Open the current URL (new tab / nouvel onglet)

    Share a precise hypothesis.

    The link restores the data version, both URLs and operations. It contains no personal browsing data.

    UNDERSTAND & REPRODUCE

    An exact calculation. A limited scope.

    What the archive documents

    188 sources with adjacency, 430 URL in total, including 242 destination-only URLs. Their outgoing links are unknown, not assumed absent.

    6729 unique directed relations, including 188 self-links retained. A loop cannot shorten a BFS path. The URL total does not represent fully audited pages.

    The 9 September archive follows the 15 August intervention. The 21 September link correction described in the article is separate evidence, absent from this historical graph.

    What the engine does

    Breadth-first search in a directed, unweighted graph. Pairs deduplicated; fragments removed; query strings retained; trailing slash removed except at the origin root. No canonical or language-based merging.

    Exact counts use BigInt. Only the first 50 paths in URL lexicographic order are materialised. Start = destination: 0 links and 1 path; no path: length unavailable, never zero.

    Calculated references, available without JavaScript

    FR · Architecture SEO : comment structurer un site pour Google et les moteurs IA

    https://www.edikka.com/ → https://www.edikka.com/insights/seo/architecture-seo-site

    In this archive, this destination can be reached from this starting URL by following at least 3 links. 37 shortest paths are available.

    1. https://www.edikka.com/
    2. https://www.edikka.com/accessibilite
    3. https://www.edikka.com/faq/seo
    4. https://www.edikka.com/insights/seo/architecture-seo-site
    path-analysis.json · graph.json

    EN · SEO architecture: how to structure a website for Google and AI engines

    https://www.edikka.com/ → https://www.edikka.com/en/insights/seo/seo-architecture-website

    In this archive, this destination can be reached from this starting URL by following at least 3 links. 2 shortest paths are available.

    1. https://www.edikka.com/
    2. https://www.edikka.com/en
    3. https://www.edikka.com/en/insights
    4. https://www.edikka.com/en/insights/seo/seo-architecture-website
    path-analysis.json · graph.json
    Provenance, original files and hashes

    Public Edikka package 1.1.0, CC BY 4.0. All eleven original files were checked byte for byte against the manifest.

    graph.json · SHA-256 388b000e10245938560fdad4cafac61c9f99edd62f1e7a924aaadc4d330da4f9

    Public import : 2026-10-01T10:05:33.315Z

    Derived display data and provenance

    pages.csv and links.csv are documentary excerpts. Only graph.json drives the general calculation. Observed H1s are identified; all other labels come from URLs.

    Replay locally with the same engine

    Node.js 22 or later, no dependencies. From a repository clone:

    npm test
npm run build
node originals/reproduce-paths.mjs
node src/replay.mjs scenario.json

    The last argument is the JSON exported by the demo. The same pure engine powers the browser, tests and this command. The calculation date does not replace the observation date.

    Full reproduction procedure