Transparency

Ownership, methods, evidence gaps and corrections belong beside the work—not in a hidden footnote.

Manifest · 2026-10-10.9 · reviewed 2026-10-10

01

Ownership & independence

Operated by Matthias Ramahi. ai-fanout.com, Crawl Foundry and SEO Fanout share an owner; those sites are connected workflows, not independent recommendations.

Crawl Foundry and SEO Fanout may later consume a user-controlled export, but shared ownership never turns either site into independent validation. ai-fanout.com must remain useful without those links.

ai-fanout.com was newly registered by Matthias Ramahi for this project. Matthias is the operator and Research Owner: he versions the tool and method, records corrections, states reviewer status and owns the EUR 25 soft / EUR 30 hard monthly cost envelope.

No independent reviewer or provider benchmark is claimed. OpenAI and Gemini Native Fanout are live behind CAPTCHA, per-bucket and site-wide quotas, an atomic monthly budget ledger and a separate public enable flag.

The server keeps no raw topic or provider response. Results are stored only when a visitor explicitly saves them in IndexedDB on that device. OpenAI observations can retain visible queries, action-linked sources and usage data. Gemini saves strip Google Grounded Results and Search Suggestions before writing query strings and run metadata locally. History is limited to 20 entries and 30 days, has visible deletion controls, is not synchronized and is not a shared response cache.

Dated public examples contain owner-run OpenAI observations only. The public OpenAI-vs-Gemini page is a comparison protocol, not a retained Gemini transcript, because Google places additional restrictions on Grounded Results and Search Suggestions.

02

Verified source register

A source supports only the statement named here. Owner records, primary documentation, standards and live technical checks remain distinct evidence roles.

  1. 2026-10-10

    Owner-run Google Sheets CSV import and native cell acceptance

    Actual selectedQueriesCsv producer with synthetic teaching strings; Sheets web File/Import/Upload, comma, conversion disabled, locale de_DE. Independent comparison confirms 110 cells and native string types. Failed automatic conversion retained. No provider call or other-application proof.

    bounded owner workflow checkOpen source
  2. 2026-10-10

    Owner-run selection and local receiver checks

    Two selected approved OpenAI strings, original en/US time/status and original 13-domain pool; versioned fragment handoff and rule-based receiver. No fresh provider call, invented source URL, query/source join, ranking or independent endorsement. Vendor rows remain documentation-only.

    bounded owner workflow checkOpen source
  3. 2026-10-10

    Profound: Query Fanout Estimator

    Prediction node with target prompt and output label; no account execution.

    documentation review onlyOpen source
  4. 2026-10-10

    Profound: dated pricing view

    Limited Trial and custom Enterprise; exact Estimator entitlement/cost remains unconfirmed.

    documentation review onlyOpen source
  5. 2026-10-10

    Ahrefs: viewing AI fanout queries

    Brand Radar supported surfaces and feature access conditions; no account or export test.

    documentation review onlyOpen source
  6. 2026-10-10

    Magganpice/fanout: extension README

    Documented manual installation, conversation access and export formats; not installed, executed or network-audited.

    documentation review onlyOpen source
  7. 2026-10-10

    Original annotated workflow, controls, prompts and export exercises

    Original diagrams and editorial annotations; only the previously approved normalized OpenAI records. Synthetic selection files use actual producers and independent CSV readback, not new provider output or Excel/Sheets import. Own dated public-page checklist makes indexing and outcomes explicitly unproven.

    owner created artifacts and verified synthetic export roundtripOpen source
  8. 2026-10-10

    Original product brief and shop question-to-page map

    Fictional two-person software buyer and storage-box assortment. Actual primary facts stay dated and bounded; tasks, paths and page assignments are editorial. No product hands-on test, cost winner, customer results or demand evidence.

    owner created fictional editorial exercisesOpen source
  9. 2026-10-10

    Google Search: AI features and your website

    Google guidance on search eligibility, established SEO practices and Web reporting; no ranking or citation guarantee.

    reviewed primary sourceOpen source
  10. 2026-10-10

    Google: managing faceted navigation URLs

    Facet URL crawling considerations and empty/invalid combinations. Our map is editorial planning, not tested configuration of a customer store.

    reviewed primary sourceOpen source
  11. 2026-10-10

    Ahrefs official pricing, reviewed 10 October 2026

    Dated displayed Lite USD monthly plan, included and additional seats. Vendor-described tasks are not hands-on quality evidence or a checkout quote.

    reviewed primary sourceOpen source
  12. 2026-10-10

    Semrush SEO pricing, reviewed 10 October 2026

    Dated displayed EUR monthly and annual SEO billing and starting additional-user price. Exact second-seat conditions, taxes and cross-currency comparison remain open.

    reviewed primary sourceOpen source
  13. 2026-10-10

    Ahrefs official terms: reuse review remains open

    Current official terms location; no general legal conclusion or permission for third-party exports/screenshots inferred.

    reviewed primary sourceOpen source
  14. 2026-10-10

    Semrush official terms: reuse review remains open

    Current official terms location; specific reuse permission remains an open task in the fictional brief.

    reviewed primary sourceOpen source
  15. 2026-10-10

    IKEA Germany SAMLA 801.029.76 manufacturer facts

    Named product dimensions 39×28×28 cm and capacity 22 litres; separately available lid. No own measurement, interior fit or general assortment claim.

    reviewed primary sourceOpen source
  16. 2026-10-10

    Original runnable offline parser and edge-case test kit

    Original educational parser code and 26 offline checks. Synthetic response fixtures are not provider output; exact strings, status states, source scope and retention exclusions are tested. No paid calls or new actual grounding data. The teaching helper explicitly differs from the live normalized wrapper.

    owner created code and synthetic fixturesOpen source
  17. 2026-10-10

    Current website export contracts and reference validation

    Draft 2020-12 reference for the current native/modelled full runs, query selections and local comparisons. Schema checked against actual service/export functions with synthetic provider stubs and semantic negative tests. No root schemaVersion is invented for full runs. Legacy planner and handoff formats remain separate.

    verified source and offline contract checksOpen source
  18. 2026-10-10

    OpenAI Responses: web-search action fields

    Supports exposed action query fields, distinct search/open/find action types and URL annotations. The documentation does not provide private consumer traces or guarantee every run exposes strings.

    verifiedOpen source
  19. 2026-10-10

    Gemini Interactions: search calls and step types

    Supports Interactions steps, google_search_call arguments and output annotations. This is a different contract from GenerateContent candidates. Synthetic teaching fixtures do not establish publication rights for actual grounded output.

    verifiedOpen source
  20. 2026-10-10

    Ajv: JSON Schema draft support

    Supports use of the Ajv2020 class for Draft 2020-12. Ajv 8.20.0 is pinned in the kit; dependency versions are locked. This source does not authenticate provider provenance or validate the truth of an exported claim.

    verifiedOpen source
  21. 2026-10-10

    Original prioritization, source-audit, measurement, repetition and reporting artifacts

    Twenty blank/worked-example files and the end-to-end fictional buying-guide workflow. Source-audit rows review public primary guidance and owner disclosure. Measurement arithmetic and attempt states are explicitly synthetic, not website results, client outcomes, provider responses or benchmarks. No paid requests or new retained grounding output.

    owner created editorial examples not provider outputOpen source
  22. 2026-10-10

    Original query-clustering, content-brief and GSC-mapping worksheets

    Blank worksheets and completed teaching examples. Seven illustrative questions and fictional website paths support the walkthrough only. No real provider output, client result, GSC measurement or keyword volume is published in these files.

    owner created editorial examples not provider outputOpen source
  23. 2026-10-10

    Google Search Console: dimensions and data groupings

    Explains query, page and filter scope and the omission of anonymised queries. Missing visible query rows do not establish zero market demand.

    verifiedOpen source
  24. 2026-10-10

    Search Console API: getting performance data

    Explains extraction and data-coverage limits. It supports a careful measurement workflow, not a claim that the tutorial or fanout ideas have measured search demand.

    verifiedOpen source
  25. 2026-10-09

    ai-fanout.com input, selection export and local-history implementation

    The input controls, separate native/modelled modes, three export scopes, query-selection CSV fields, and browser-local history are described from the site's actual implementation. The planning prompt and input examples are original editorial instructions, not provider output or measured keyword demand.

    verified source and regression contractOpen source
  26. 2026-10-09

    MDN: IndexedDB API

    Explains browser-side structured storage and its origin boundary. The site's twenty-entry and thirty-day limits come from its own implementation, not from a universal browser quota.

    verifiedOpen source
  27. 2026-10-09

    Introduction to Information Retrieval: query expansion

    Supports the vocabulary-expansion distinction. It does not document the internals or exact queries of a current consumer AI search session.

    verifiedOpen source
  28. 2026-10-09

    Owner full-inventory link and reader-path review

    A complete 35-URL same-origin sitemap crawl found no orphan URLs, but guide chapters did not directly connect to their dated proof pages and example bodies offered downloads without interpretive guide paths. Added tables and reproduction instructions are owner-created editorial material; published provider fixtures and observation dates are unchanged.

    verified http and html not ranking predictionLocal or central record
  29. 2026-10-09

    Owner Search Console and Crawl Foundry SEO review

    Prioritises existing US generator intent and provider guides using authenticated GSC data. Supports editorial prioritisation only; not a promise of rankings, indexing or clicks. Public copy is owner-created and contains no newly published provider output.

    verified measurement not ranking predictionLocal or central record
  30. 2026-08-22

    Canonical ai-fanout.com domain dossier

    Confirmed ownership and the owner-approved modelled-fanout tool plus bilingual learning-hub role.

    acceptedLocal or central record
  31. 2026-08-22

    Canonical owned-domain registry entry

    Records ai-fanout.com as owned and newly purchased for the current project.

    verifiedLocal or central record
  32. 2026-08-22

    Matthias Ramahi imprint

    Operator name, postal address, contact email and editorial responsibility used for the ai-fanout.com legal notice.

    verifiedOpen source
  33. 2026-08-22

    Canonical operator legal source

    Repository source confirming Matthias Ramahi, Kempener Straße 44, 40699 Erkrath, info@matthiasramahi.de and responsibility under section 18(2) MStV.

    verifiedLocal or central record
  34. 2026-08-22

    Vercel Privacy Notice

    Hosting-provider disclosures about end-user request data, retention, providers and international transfers used in the site-specific privacy notice.

    verifiedOpen source
  35. 2026-08-29

    Umami tracker configuration

    Documents domain restriction, Do Not Track support, exclusion of URL search and hash values, and performance measurement used by the site tracker.

    verified liveOpen source
  36. 2026-08-29

    Umami tracker functions, event data and sessions

    Documents custom event calls and session derivation. The implementation adds a strict local allowlist that excludes raw tool inputs, outputs, source addresses and user identifiers.

    verified liveOpen source
  37. 2026-08-22

    AI features and your website

    Google describes query fan-out publicly without exposing a general private-query inspection interface.

    verifiedOpen source
  38. 2026-08-31

    AI Mode in Google Search: updates from Google I/O 2025

    Google states that AI Mode uses query fan-out to break a question into subtopics and issue multiple searches; it does not provide ai-fanout.com with a private AI Mode query trace.

    verified liveOpen source
  39. 2026-10-10

    Optimizing for generative AI features on Google Search

    Normal SEO foundations remain relevant; pages made primarily for fan-out variations are not the accepted content model.

    verifiedOpen source
  40. 2026-08-22

    Creating helpful, reliable, people-first content

    Supports original, substantial, audience-first content with transparent sourcing and production context.

    verifiedOpen source
  41. 2026-08-22

    Spam policies for Google web search

    Defines scaled content abuse regardless of production method.

    verifiedOpen source
  42. 2026-08-22

    NIST AI RMF Generative AI Profile

    Supports explicit measurement, documentation, monitoring and limitations for generative-AI evaluations.

    verifiedOpen source
  43. 2026-08-22

    PROV-O: The PROV Ontology

    Provides provenance concepts for entities, activities, agents, derivations, sources and versions.

    verifiedOpen source
  44. 2026-08-22

    RFC 3339: Date and Time on the Internet

    Supports an interoperable timestamp representation tied to UTC.

    verifiedOpen source
  45. 2026-08-22

    The FAIR Data Principles

    Supports reusable research data with metadata, provenance and clear usage licenses.

    verifiedOpen source
  46. 2026-08-22

    RDAP record for ai-fanout.com

    Registration event dated 2026-08-16 and current registry state.

    verifiedOpen source
  47. 2026-08-22

    Current custom-domain response

    Custom domain serves canonical indexable 200 pages without meta or header noindex, a crawlable robots file, automatic sitemap, intentional tracker noindex and real 404 responses.

    verified indexable liveOpen source
  48. 2026-08-22

    Current Vercel project alias

    The Vercel project alias serves the same deployment with canonical URLs pointing to the custom domain and no global noindex.

    verified indexable previewOpen source
  49. 2026-08-22

    Browser-local tool implementation

    Historical local comparison utilities contained no transmission or browser storage path. Their public routes were retired on 2026-08-24 and now return 404.

    verifiedLocal or central record
  50. 2026-08-22

    Public source repository

    Current code provenance and absence of a repository license file.

    verifiedOpen source
  51. 2026-08-22

    Dated Evidence Lab synthetic fixture

    Exact synthetic inputs, method version and deterministic expected results for the local calculation workflow; not an independent provider benchmark.

    owner created reproducibleLocal or central record
  52. 2026-08-22

    Astro sitemap integration

    Automatic sitemap generation from statically generated Astro routes with route filtering.

    verifiedOpen source
  53. 2026-08-26

    OpenRouter model catalog

    Confirms the pinned GPT-5.6 Luna and Gemini 3.7 Flash OpenRouter model IDs and structured-output support used by the secondary Search Ideas allowlist.

    verified liveOpen source
  54. 2026-08-22

    OpenAI Responses API reference

    Supports one stateless Responses API call, store false, max output tokens and strict JSON Schema output.

    verifiedOpen source
  55. 2026-10-10

    OpenAI web search tool guide

    Documents web_search_call output and search actions that usually, but not always, include the query or queries searched. Search actions, exposed strings and cited sources therefore need separate interpretation. It does not expose chain of thought or guarantee that every route returns query strings. Rechecked for the original 2026-10-10 zero-query decision table; no new provider observation is published.

    verifiedOpen source
  56. 2026-08-31

    Searching the web with ChatGPT

    OpenAI explains that ChatGPT Search can rewrite a question into one or more targeted queries and can issue more specific follow-up searches; this description does not expose a private ChatGPT search history to ai-fanout.com.

    verified liveOpen source
  57. 2026-08-26

    GPT-5.6 Luna model documentation

    Confirms the direct API model ID gpt-5.6-luna and Responses API web-search support.

    verifiedOpen source
  58. 2026-08-28

    GPT-5.6 Luna model pricing

    Lists USD 0.20 per million input tokens and USD 1.20 per million output tokens used by the dated per-run estimate.

    verifiedOpen source
  59. 2026-08-27

    OpenAI web-search pricing

    Lists the standard web_search tool at USD 10 per 1,000 calls plus search-content tokens at model rates.

    verifiedOpen source
  60. 2026-08-26

    Gemini grounding with Google Search

    Documents Gemini 3.7 Flash Google Search, google_search_call arguments.queries, cited URL annotations and per-query billing.

    verifiedOpen source
  61. 2026-08-27

    Gemini Developer API zero data retention

    Documents Interactions API store false and the unavoidable 30-day provider retention for prompts, context and output when Grounding with Google Search is used.

    verifiedOpen source
  62. 2026-08-27

    Gemini Developer API pricing

    Lists Gemini 3.7 Flash token prices through 2026 and the 5,000 monthly Search-grounding allowance followed by USD 14 per 1,000 search queries.

    verifiedOpen source
  63. 2026-08-27

    Gemini API Additional Terms

    Restricts caching, syndication, analysis and reuse of Google Grounded Results and Search Suggestions; drives the Gemini local-save redaction and public-example boundary.

    verified restrictiveOpen source
  64. 2026-08-24

    OpenRouter web search server tool

    Documents native web search, location hints and use limits. Exact query-string preservation for the chosen OpenAI route remains NOT PROVEN until a live response fixture is inspected.

    verified with open gateOpen source
  65. 2026-08-22

    GPT-5.4 nano model documentation

    Historical planner model reference. It is not the current observed-query route.

    superseded not currentOpen source
  66. 2026-08-22

    Cloudflare Turnstile documentation and privacy policy

    Server-side human-verification boundary and required privacy disclosure.

    verifiedOpen source
  67. 2026-08-22

    Modelled AI fanout tool secured release candidate

    Strict input and ten-item output schemas, one selected allowlisted OpenRouter model call, atomic quota and budget ledger, timeout, no raw server archive and fail-closed environment gate.

    verified closed by defaultLocal or central record
  68. 2026-08-26

    Direct OpenAI and Gemini native-fanout implementation

    One direct provider request, strict provider allowlist, provider-exposed query/source parsing, honest zero-query output, CAPTCHA-first shared atomic quota and budget accounting, no raw server archive and separate public flag. Bounded live production fixtures passed for OpenAI and Gemini on 2026-08-26.

    verified liveLocal or central record
  69. 2026-08-27

    Opt-in browser-local result history

    Explicit save action, IndexedDB-only storage, maximum 20 entries, 30-day expiry, individual deletion, clear-all and local JSON/CSV export without accounts or shared response caching.

    verified local onlyLocal or central record
  70. 2026-08-27

    Dated OpenAI fanout example observations

    Four owner-run OpenAI API observations record exact inputs, timestamps, exposed query strings, search-action source domains, usage, method versions and response status. Versioned JSON, normalized CSV and a JSON Schema are published. They are not an independent benchmark.

    verified reproducibleOpen source
  71. 2026-08-28

    Bounded monthly observation plan

    Defines 12 repeat topics, fixed run controls, ownership and publication boundaries. It is a plan and contains no claimed monthly results.

    planned not resultsOpen source
  72. 2026-10-10

    Current OpenRouter model catalog

    Live exact IDs openai/gpt-6-luna, google/gemini-3.8-flash and anthropic/claude-haiku-5.5; structured outputs and per-model reasoning controls. No inference output retained or published.

    verified liveOpen source
  73. 2026-10-10

    GPT-6 Luna model and pricing

    Direct model ID gpt-6-luna, Responses web-search support, USD 0.10/M input and USD 0.50/M output for short prompts.

    verified liveOpen source
  74. 2026-10-10

    Gemini 3.8 Flash pricing

    USD 0.75/M input and 3.75/M output through 2026-12-31, 1.50/M and 7.50/M from 2027-01-01; shared 5,000 monthly Search queries then USD 0.014/query. Range preserves unknown allowance.

    verified liveOpen source
  75. 2026-10-10

    Claude Haiku 5.5 release and pricing

    USD 0.10/M input and 0.50/M output up to 100k prompt tokens, 0.50/M and 2.50/M above that threshold. This is a list-price reference, not a benchmark.

    verified liveOpen source
  76. 2026-10-10

    Anthropic native web search contract

    Anthropic native-search contract and vendor-specific tool/result boundaries. Current routed runs parse OpenRouter Responses action fields; Messages query inputs are not assumed to survive routing.

    verified liveOpen source
  77. 2026-10-10

    Original runnable offline parser and edge-case test kit

    Original educational parser code and 27 offline checks. Synthetic response fixtures are not provider output; exact strings, status states, source scope and retention exclusions are tested. No paid calls or new actual grounding data. The teaching helper explicitly differs from the live normalized wrapper.

    owner created code and synthetic fixturesOpen source
  78. 2026-10-10

    Current website export contracts and reference validation

    Draft 2020-12 reference for the current native/modelled full runs, query selections and local comparisons. Schema checked against actual service/export functions with synthetic provider stubs and semantic negative tests. No root schemaVersion is invented for full runs. Legacy planner and handoff formats remain separate.

    verified source and offline contract checksOpen source
  79. 2026-10-10

    OpenRouter native server tools and Responses routing evidence

    Responses and Chat Completions native server tools, model-only provider routing and native-engine pipeline evidence. Gemini Chat exposes request counts and run-level citations, but no query strings in the checked route. External engine responses fail closed. Haiku Messages loses query inputs in the checked routed response, so current native runs use Responses action fields instead.

    verified documentation and protected runtime metadataOpen source
03

Rights decisions

Current repository code and copy
owner created no open license

The public repository does not grant reuse rights unless a later LICENSE file does so explicitly.

Third-party screenshots, copied text and stored model outputs
provenance and rights required before publication

Publication requires verified provenance, clear attribution and the necessary provider, copyright and privacy permissions.

AI provider outputs and source lists used in a future site-run dataset
provider specific publication boundary

The live endpoint makes one transient provider request and keeps no raw server archive. Four reviewed OpenAI observations are published as a normalized owner-run fixture under OpenAI's customer-output rights. Google Grounded Results and Search Suggestions are not stored or published; Gemini local saves strip them before retaining query strings and run metadata.

Typography and production imagery
system fonts and owner commissioned ai images

The website uses system font fallbacks. Five text-free guide illustrations were commissioned by the owner through Higgsfield on 2026-08-27 and generated with nano_banana_2 after the requested model was adjusted by the service. Four additional topic-specific guide illustrations were commissioned directly by the owner through Google's gemini-3-pro-image model at native 2K on 2026-08-31; their prompts and generation metadata are stored with the project documentation. All nine illustrations are delivered locally as WebP files under public/images/guides and contain no copied third-party logos, screenshots or artwork.

External standards and documentation
linked and paraphrased

Reference pages link to primary sources and paraphrase narrowly supported facts without republishing source documents or protected artwork.

User topic, selected mode/provider, language, country and transient fanout result
server transient optional browser local copy

Input is transmitted to the ai-fanout.com endpoint and routed either directly to OpenAI or Google for native search, or through OpenRouter for modelled search ideas. The server stores no raw input or provider response. Only after an explicit save action may the visible result be retained in IndexedDB on that device for up to 30 days; it is not synchronized or reused as a shared cache.

Umami page and product analytics
privacy minimized first party measurement

The owner-operated Umami endpoint receives page analytics and only allowlisted product-event categories and bounded counts. URL search/hash values, raw keywords, provider outputs, query strings, source addresses, local-history content and user identifiers are excluded.

04

Current route policy

This is a fresh registration. The route manifest describes only the current product and normal unknown-path handling.

Sitemap source
central allowlist with editorial lastmod and hreflang
Excluded routes
/404, /tracker, /api, /lab, /protocol-builder, /research, /research/methodology, /datasets, /protocols/example-2026-08-22
Unknown paths
404
Catch-all redirect
Disabled

Canonical English and German 200 pages are indexable and discovered through an automatically generated sitemap. The operational tracker remains noindex. Retired comparison and research routes, and all other unknown paths, return a real 404.

05

Still unknown

  • Exact billed cost per OpenAI and Gemini run remains subject to provider-dashboard review; the UI shows a dated list-price estimate and the ledger retains a conservative EUR 0.10 reservation per Native Fanout request
  • Provider terms for transient processing across OpenRouter, OpenAI and Google require ongoing operator review
  • No public Gemini Grounded Result corpus or independent provider benchmark is authorized; the OpenAI-vs-Gemini page remains a local comparison protocol
  • Google Search Console post-change crawl and index state after an eventual safe production release
06

Corrections & contact

For a factual, rights or methodology correction, email info@matthiasramahi.de or open a public issue in the source repository and name the affected route and evidence. Do not post vulnerabilities, credentials, private endpoints or personal data publicly.

Open the source repository

Security concerns use GitHub private vulnerability reporting as described in SECURITY.md.