--- optiview_delivery: edge_optimized content_type: semantic_markdown source_url: https://optiview.ai/proof/methodology/ brand_content: included publication_version: "9.0" published_at: 2026-10-16T20:36:07Z composition_version: composition-v2.1.0 presentation_version: presentation-v1.2.3 brand_content_source: narrative_graph delivery_publication: graph:26 narrative_versions: - 26 narrative_version: 26 --- # How we check what happened. Full study outcomes, source comparisons and delivery records behind the observations. The Optiview Experiment · Technical appendix [Return to the experiments →](https://optiview.ai/proof/) Three layers of the investigation ## Follow the information from publication to an answer. Each layer answers a different question. A verified delivery does not establish retrieval, and retrieval does not establish an effect on an answer. ### Publishing and delivery What information did Optiview actually serve? [Routing audit →](https://optiview.ai/proof/methodology/#search-policy) [Bot identities and delivery records →](https://optiview.ai/proof/methodology/#bot-delivery) ### Retrieval and accessibility What information did the browsing system actually receive? [Local transformation study →](https://optiview.ai/proof/methodology/#browser-reference) [Completed Exa and You.com follow-up →](https://optiview.ai/proof/methodology/#live-provider-method) ### Interpretation and outcomes What did AI say, and can we attribute a difference? [Retained conversations →](https://optiview.ai/proof/#live-evidence) [What remains unproven →](https://optiview.ai/proof/#evaluations) Search-crawler delivery audit · October 29, 2026 ## Historical delivery and current policy are different. The retained activity includes one prepared HTML response labeled Googlebot and 19 labeled bingbot, all on October 21 under parser semantic-v3.10.12. We keep those historical records; they do not describe today’s routing. Current policy passes traditional search crawlers through to the original response. Operator checks across four connected websites, using both crawler names and three Accept formats, returned the original web pages. These declared-name checks test routing, not authenticated visits from Google or Microsoft. Recent recorded requests for both names also show no optimization. Googlebot activity does not establish Gemini use, and bingbot activity does not establish use in an AI answer. A response’s format alone does not establish that Optiview prepared it: original files can themselves be Markdown or JSON. [Inspect the retained policy audit and 28 operator checks (JSON)](https://optiview.ai/proof/search-policy-20261029.json) Open-source reference code · Optiview-run pilot ## Checking what survives between a website and a model. We use OpenAI’s published gpt-oss browser code as a reference for testing how website content becomes model-readable text. We execute its pinned HTML-to-text processing and browser display-window functions against retained website captures. The code is a reference for this experiment; it is not our production delivery engine. ### What we ran Two sites, eight fixed questions, three source conditions and three repetitions: 72 answers using Cloudflare-hosted gpt-oss-120b. Ordinary HTML, prepared content without the graph and prepared content with the published graph all pass through the same reference HTML-to-text processing. The answer experiment receives the full transformed text; initial-window visibility is measured separately. The no-graph condition removes only the published narrative block from the captured composed source. It is an offline ablation, not a separately deployed website. Added context supplies information absent from the baseline; answer completeness on those details cannot establish a format advantage. Shared facts are evaluated separately. ### Why we are not claiming an answer-quality improvement We retained all 72 answers and 71 usable automated evaluations. Manual checks found that the evaluator sometimes penalized a correct statement that information was absent from the supplied source, or confused missing details with errors. The raw scores remain available as diagnostics; they do not establish improved accuracy. All 18 responses to deliberately unanswerable questions correctly abstained. ### What the code lets us inspect We retain the input, transformed text, resolved links and the initial display produced with the reference browser’s default 1,024-token budget. In this local test, all 13 registered graph-context supporting passages survived the full transformation; none appeared in the initial window. Additional navigation can expose later text. We did not measure an agent’s choice to navigate. ### What the controlled pilot did not test The reference implementation supports Exa and You.com backends. Both live provider paths were NOT TESTED in this local pilot because credentials were unavailable. A separate live follow-up is reported below. Our synthetic adapter checks test request construction and response handling; they do not establish live compatibility or that those providers retrieve Optiview’s published context. This is an Optiview-run experiment using OpenAI-designed open-source code. It is not an OpenAI-conducted study, endorsement or partnership, and does not establish how production ChatGPT browses. It does not measure organic discovery, citation lift or recommendation frequency. ### Inspect the pinned implementation and our records - [Browser navigation and display code](https://github.com/openai/gpt-oss/blob/5681a6b6552635e7b29cde94a61ca6a7aee90ebd/gpt_oss/tools/simple_browser/simple_browser_tool.py) - [HTML-to-text processing code](https://github.com/openai/gpt-oss/blob/5681a6b6552635e7b29cde94a61ca6a7aee90ebd/gpt_oss/tools/simple_browser/page_contents.py) - [Exa and You.com backend code](https://github.com/openai/gpt-oss/blob/5681a6b6552635e7b29cde94a61ca6a7aee90ebd/gpt_oss/tools/simple_browser/backend.py) - [Frozen protocol and scoring plan](https://optiview.ai/proof/browser-study-20261029/protocol.md) - [All answers, scores and conditions](https://optiview.ai/proof/browser-study-20261029/results.json) - [Scoring limits and manual disagreements](https://optiview.ai/proof/browser-study-20261029/scoring-report.md) - [Source and window comparisons](https://optiview.ai/proof/browser-study-20261029/context-survival.json) - [Sources, code and reproduction packet](https://optiview.ai/proof/browser-study-20261029/reproduction.zip) Separate live retrieval follow-up · October 29, 2026 ## Publishing context is one step. Getting it back is another. Our local test showed that graph context can survive the browser’s transformation when supplied. We then tested what the same pinned OpenAI open-source browser’s Exa and You.com retrieval paths actually returned. ### 12 real requests succeeded. Two websites. Four direct URL requests and eight fixed, unbranded searches. All returned HTTP 200, without retries or query changes. ### The complete graph passages weren’t supplied. All four direct responses returned page content, but none contained a complete registered graph supporting passage. Separate checks confirmed all 13 passages in the current prepared website representations. ### One search returned a target website. Exa returned Learnings.org at position 13 for a question about free corporate buzzword tools. It was below the browser’s initial viewing window, but accessible through further navigation. Some ordinary-page meaning overlapped with the published context: 8 of 26 provider-by-passage checks found partial related information. That is not evidence that the information came from the graph. These observations separate what a website publishes, what a provider returns and what a browsing model initially sees. ### Direct retrieval and search, separately | Provider | Direct opens | Complete passages: Rhythm90 / Learnings | Target in 20 search results | | --- | --- | --- | --- | | Exa | 2 of 2 returned content | 0 of 7 / 0 of 6 | 1 of 4 searches | | You.com | 2 of 2 returned content | 0 of 7 / 0 of 6 | 0 of 4 searches | The complete passages were absent at the observable processor input, rather than lost only during transformation or truncation. Exa labeled the Rhythm90 result “crawled” and Learnings “cached”; these are provider reports, not independent freshness verification. You.com supplied no freshness metadata. We cannot determine whether representation selection, caching, extraction or another retrieval difference explains the results. All 160 search positions are retained. The other seven searches did not return the target hostname; that does not mean the sites are unindexed. The Exa target listing begins at line 228, outside the initial window ending before line 55. Its link and the pinned find operation expose it later. No model chose to navigate or fetch that result. Exa’s search listing uses provider-generated summaries; You.com’s uses snippets. Neither is treated as full-page graph preservation or an independently verified assistant recommendation. ### Reproduce the observable processing Offline replay reproduced all 12 retained responses’ browser text, links, windows and find behavior. Full raw search responses remain private because they include unrelated third-party content. The public packet includes all result positions, four complete first-party direct responses and the selected target search result; its offline replay covers the four complete direct responses. Credentials and account data are excluded. Ten events in the origin-and-study window were conservatively excluded from cumulative marketing activity; timing does not establish which provider caused a visit. - [Frozen protocol](https://optiview.ai/proof/live-provider-20261029/protocol.md) - [All 12 registered cases](https://optiview.ai/proof/live-provider-20261029/cases.json) - [Results, search positions and source checks](https://optiview.ai/proof/live-provider-20261029/results.json) - [Semantic passage audit](https://optiview.ai/proof/live-provider-20261029/manual-audit.json) - [Findings and limitations](https://optiview.ai/proof/live-provider-20261029/findings.md) - [Rank-13 target and browser navigation](https://optiview.ai/proof/live-provider-20261029/cases/exa-learnings-discovery.json) - [Public evidence and offline replay](https://optiview.ai/proof/live-provider-20261029/reproduction.zip) These 12 requests do not establish why complete graph context was absent, how production ChatGPT browses, whether a model would navigate further, or whether Optiview improves citations or recommendations. This Optiview-run use of OpenAI-designed reference code is not an endorsement or partnership. How we run the tests ## Keep the question, the answer and the evidence. These are first-party observations from selected questions, not a representative benchmark. We retain outcomes beyond the examples featured on the experiment page, including omissions, technical interruptions and protocol deviations. 1. **Register the question** Record the question, platform and product mode. Distinguish questions that name no business from checks that supply a website URL. 2. **Capture the answer** Use new temporary or incognito conversations where available. Retain the wording and cited sources. A usable answer is not retried to obtain a preferred result. 3. **Compare with the source** Look for published facts, qualifications and exact narrative phrases. Check ordinary pages too, and compare available response hashes and publication versions. Inspect every study’s results and retained records ### Separate cohorts. No combined performance score. “Usable answers” excludes technical interruptions. “Omissions” means a completed answer did not mention the target website. “Inconclusive” includes unfinished or blocked combinations. An appearance alone does not establish source-language use. Retained evaluation coverage · October 2026 | Question set | Planned | Usable answers | Appearances | Omissions | Inconclusive | Records | | --- | --- | --- | --- | --- | --- | --- | | Oct 22 · initial unbranded questions | 48 | 47 | 11 | 36 | 1 | [Inspect](https://optiview.ai/proof/experiment-20261029/earlier-outcomes.json) | | Oct 22 · search-enabled questions | 16 | 16 | 8 | 8 | 0 | [Inspect](https://optiview.ai/proof/experiment-20261029/earlier-outcomes.json) | | Oct 23 · new unbranded questions | 48 | 48 | 2 | 46 | 0 | [Inspect](https://optiview.ai/proof/discovery-2026-10-23/outcomes.json) | | Oct 23 · exploratory replications | 20 | 19 | 3 | 16 | 1 | [Inspect](https://optiview.ai/proof/discovery-2026-10-23/outcomes.json) | | Oct 24 · additional unbranded questions | 32 | 32 | 1 | 31 | 0 | [Inspect](https://optiview.ai/proof/discovery-2026-10-24/outcomes.json) | | Oct 25 · source-language questions | 64 | 45 | 10 | 35 | 19 | [Inspect](https://optiview.ai/proof/discovery-2026-10-25/) | | Oct 26 · Gemini continuation | 12 | 3 | 0 | 3 | 9 | [Inspect](https://optiview.ai/proof/gemini-followup-2026-10-26/) | | Oct 28 · aligned unbranded questions | 8 | 8 | 2 | 6 | 0 | [Inspect](https://optiview.ai/proof/refresh-20261028/outcomes.json) | | Oct 29 · Unbranded specific-need questions | 4 | 3 | 0 | 3 | 1 | [Inspect](https://optiview.ai/proof/experiment-20261029/outcomes.json) | | Oct 29 · Directed URL source-use checks | 4 | 4 | 4 | 0 | 0 | [Inspect](https://optiview.ai/proof/experiment-20261029/outcomes.json) | The October 26 continuation revisits unfinished questions from October 25; those rows overlap and must not be added. Question sets, modes and publication baselines differ across studies. None is a controlled before-and-after comparison. ### Modes and exceptions Recorded modes include ChatGPT Medium effort, Claude Sonnet 5.5 Medium, Perplexity Search and Gemini Pro. Later ChatGPT conversations explicitly used Unpersonalized temporary mode. The Gemini continuation also used 3.8 Flash for some unfinished questions. Temporary modes do not independently establish signed-out anonymity or the absence of every account influence. The October 23 study retained one identical retry after a technical failure and an “Answer now” intervention on a stalled Gemini replication. Subsequent protocols distinguished valid omissions from technical retry eligibility. The October 29 first Gemini temporary conversation reset after a second temporary tab opened; it remains inconclusive. Each linked record retains the applicable method and exceptions. Private account navigation is excluded from public captures. Selected excerpts and source comparisons are not independent third-party attestations. Existing proof archives and search caches can themselves carry the published wording. Inspect how source wording and delivery are checked ### Language and meaning are different checks. For the latest graph-language checks, exact sequences of at least eight words are compared after Unicode, case and punctuation normalization. Paraphrases are reviewed separately. We check whether the wording was in the question or captured ordinary pages before calling it distinctive graph evidence. A delivery hash identifies retained response bytes or a deterministic reconstruction of prepared HTML. Provider-IP verification is stronger than a declared bot name; Cloudflare category verification does not independently confirm a named provider. Request timing corroborates an observation, but does not uniquely tie a retrieval to one answer. Google-origin traffic is not attributed to Gemini from timing alone. Source content, publication versions and responses can be inspected. Passing checks does not certify that the website supplied every fact. Before-and-after movement alone does not establish that Optiview caused it. A clearer source does not guarantee inclusion or better outcomes. - [Latest graph-language conversations, source passages and hashes](https://optiview.ai/proof/graph-language-20261029/index.json) - [Original directed retrieval evidence packet](https://optiview.ai/proof/evidence-2026-10-22/index.json) - [Unbranded discovery source comparisons](https://optiview.ai/proof/discovery-2026-10-23/index.json) REVIEWED EVIDENCE: October 22, 2026 **Single-page parser:** semantic-v3.10.14. The deterministic parser includes source-owned title and editorial metadata repairs, responsive heading handling, interface heading exclusion, punctuation serialization, source-backed appendix labels, and valid Markdown for linked cards. Core parsing does not generate FAQs, infer company-wide information or use an LLM. Existing source-review bindings, custom selectors and approved published supplements retain their review controls. ## Parser validation An internal engineering check on a fixed set of inputs. The 660 executions include reference inputs, deliberate variations, controls and 100 captured real pages. Zero unexplained differences means the checked outputs matched the accepted baseline; it is not a 100% extraction-accuracy score. These tests do not measure how much content a new or complex JavaScript-heavy page may omit, or whether AI answers improve. **Source executions** 660 **Unexplained differences** 0 **Composed outputs** 65 **Unchanged extraction abstentions** 21 660 frozen-source serializer comparisons under the repaired parser, plus exact output comparisons on 1,000 core sources and 660 retained inputs against the accepted engineering candidate. Generated labels are distinguished from source prose. These retained-source checks do not establish 99.9% accuracy or AI outcomes. Extraction `semantic-v3.10.14`Composition `composition-v2.1.0` Corpus, checks and scope **Corpus:** 164 gold inputs, 378 mutations, 18 meaningful controls and 100 captured real pages. * Serializer equivalence under the repaired parser, with separate frozen-output comparisons against the accepted engineering candidate. * Literal code payload preservation, source-prose checks and source inventory comparison. * Destination preservation and deterministic composition. * Adversarial cases covering distinct topics, unequal memberships and ambiguous directory spans. These versions identify the source-code baseline reviewed for this site. For an individual result, inspect its returned versions and source mode. A static website label does not attest to every deployed environment. - [Download the validation record](https://optiview.ai/parser-validation.json) - [Inspect your own page](https://optiview.ai/demo/) ### Bot activity, response formats and compatibility checks Production activity is separate from controlled tests. Known operator probes and registered study traffic are excluded from the cumulative activity evidence. Counts include repeated requests, not unique people or independently authenticated providers. Real delivery activity ## AI bots request it. Optiview serves it. Tens of thousands of requests carrying bot names. Thousands of prepared content responses served. Consistent daily AI-bot activity. Optiview delivers clean, readable content as Markdown and prepared HTML, with published Narrative Graph context where applicable. **40,000+**requests with bot names observed **Thousands**prepared content responses served **80+**bot and crawler names observed **Daily**consistent AI-bot activity **Clean Markdown.** Readable source content, headings and links that compatible AI tools can request directly. **Prepared HTML.** Optiview’s prepared source in HTML for compatible AI requests, including applicable published graph context. Delivery tested end to end ### 44 of 44 live compatibility checks passed Across 4 connected websites, operator checks verified ordinary HTML, explicit Markdown requests, prepared HTML for declared AI-tool requests and original-page delivery for a traditional search crawler. Prepared-source comparisons retained the published graph context. These are synthetic compatibility checks, not visits from real AI providers. HTML source comparisons remove only Cloudflare’s injected analytics beacon; retained raw response hashes remain available. [Inspect the request matrix and response fingerprints](https://optiview.ai/proof/refresh-20261028/delivery-checks.json) These names appear in our website request records. The directory explains their roles and distinguishes recorded prepared-content delivery from general website activity. Search crawlers, AI search tools and training crawlers do different jobs. Meta · AI crawler ### meta-externalagent Meta’s automated web crawler, classified as an AI Crawler in Cloudflare’s bot directory. Prepared content servedMarkdown · Prepared HTML Bot-category evidence recorded [About meta-externalagent](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) Amazon · AI crawler ### Amazonbot Amazon’s crawler gathers information to improve its products and services; Amazon says this content may also be used to train its AI models. Prepared content servedMarkdown Bot-category evidence recorded [About Amazonbot](https://developer.amazon.com/amazonbot) Perplexity · AI search ### PerplexityBot Crawls websites for Perplexity search results. Perplexity distinguishes this search crawler from foundation-model training. Prepared content servedMarkdown · Prepared HTML Provider-IP matches recorded [About PerplexityBot](https://docs.perplexity.ai/docs/resources/perplexity-crawlers) Anthropic · AI search ### Claude-SearchBot Collects and analyzes web content to improve the relevance and accuracy of Claude’s search responses. Prepared content servedMarkdown Bot-category evidence recorded [About Claude-SearchBot](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) Anthropic · AI crawler ### ClaudeBot Collects public web content that may contribute to training Anthropic’s models. It is separate from Claude’s search and user-requested retrieval bots. Prepared content servedMarkdown Bot-category evidence recorded [About ClaudeBot](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) OpenAI · User-requested retrieval ### ChatGPT-User May fetch a page in response to an action in ChatGPT or a custom GPT. It is separate from OpenAI’s automatic search crawler. Prepared content servedMarkdown · Prepared HTML Provider-IP matches recorded [About ChatGPT-User](https://developers.openai.com/api/docs/bots) DuckDuckGo · AI-assisted answers ### DuckAssistBot Fetches pages for DuckDuckGo’s AI-assisted answers. DuckDuckGo says this data is not used to train AI models. Prepared content servedMarkdown · Prepared HTML Bot-category evidence recorded [About DuckAssistBot](https://duckduckgo.com/duckduckgo-help-pages/results/duckassistbot) OpenAI · AI crawler ### GPTBot Crawls content that may be used to train OpenAI’s generative AI models. A request does not establish that the content entered training. Prepared content servedMarkdown Provider-IP matches recorded [About GPTBot](https://developers.openai.com/api/docs/bots) Apple · Search and discovery ### Applebot Apple’s web crawler supports search and discovery features, including experiences in Spotlight, Siri and Safari. Prepared content servedPrepared HTML Bot-category evidence recorded [About Applebot](https://support.apple.com/en-us/119829) OpenAI · AI search ### OAI-SearchBot OpenAI’s automatic search crawler helps surface websites in ChatGPT search. Prepared content servedMarkdown Provider-IP matches recorded [About OAI-SearchBot](https://developers.openai.com/api/docs/bots) Anthropic · User-requested retrieval ### Claude-User Fetches web content in response to user requests in Claude. Prepared content servedPrepared HTML Observed request name; identity unverified [About Claude-User](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) Perplexity · User-requested retrieval ### Perplexity-User May fetch a page to help answer a user’s question in Perplexity. Website requests observed Observed request name; identity unverified [About Perplexity-User](https://docs.perplexity.ai/docs/resources/perplexity-crawlers) Meta · AI assistant ### meta-externalfetcher Classified by Cloudflare as Meta’s AI Assistant fetcher, separate from its automated crawler. Website requests observed Observed request name; identity unverified [About meta-externalfetcher](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) Google · AI crawler ### Google-CloudVertexBot A Google crawler associated with Vertex AI; Cloudflare classifies it as an AI Crawler. Its name does not establish Gemini conversation use. Prepared content servedMarkdown Observed request name; identity unverified [About Google-CloudVertexBot](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) Google · Search crawler ### Googlebot Google’s search crawler. Ordinary search crawling is separate from evidence of use in an AI answer. Original-response policy Historical records include prepared HTML. Current policy passes through the original response; search crawling does not establish AI use. [Inspect the policy audit](https://optiview.ai/proof/methodology/#search-policy) Provider-IP matches recorded [About Googlebot](https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers) Google · Other crawling ### GoogleOther Google’s general-purpose crawler for uses outside its main Search crawler. Website requests observed Bot-category evidence recorded [About GoogleOther](https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers) Microsoft · Search crawler ### bingbot Microsoft’s Bing search crawler, classified separately from AI crawlers in Cloudflare’s bot directory. Original-response policy Historical records include prepared HTML. Current policy passes through the original response; search crawling does not establish AI use. [Inspect the policy audit](https://optiview.ai/proof/methodology/#search-policy) Bot-category evidence recorded [About bingbot](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) Common Crawl · Open web crawling ### CCBot Common Crawl’s crawler contributes to its open repository of web crawl data. A request does not establish downstream model use. Prepared content servedMarkdown Observed request name; identity unverified [About CCBot](https://commoncrawl.org/ccbot) Declared crawler · Observed request name ### YouBot Requests using this bot name appear in our delivery records. The name alone does not independently confirm a provider. Website requests observed Observed request name; identity unverified All observed bot and crawler names The complete retained name inventory includes search, SEO, social, research and other automated request labels as well as AI tools. Labels alone do not independently confirm an operator. A request can be recorded even when it failed or received the original website. * adsbot * agentcensus-crawler * ai2bot * Amazonbot * apcosysbot * Applebot * argorantbot * arnorbot * awariobot * baiduspider * bingbot * bitmoversbot * bitsightbot * bravebot * bytespider * CCBot * cgmimm-bot * chatglm-spider * ChatGPT-User * cl0q-crawler * Claude-SearchBot * Claude-User * ClaudeBot * databot * deepseekbot * dnbcrawler * domains-crawler * donoratlasbot * dotbot * DuckAssistBot * duckduckbot * elevenlabscrawler * everytabbot * exasearchbot * firmablebot * Google-CloudVertexBot * Googlebot * GoogleOther * GPTBot * grokbot * halobot * hostio-bot * iaskspider * keenablebot * kimi-searchbot * kimibot * leadbot * leadoceanbot * linkedinbot * lohisoftbot * mailsiftbot * meta-externalagent * meta-externalfetcher * mj12bot * mochiimagecrawler * moonshotbot * novantrodomainsignalcrawler * OAI-SearchBot * oi-crawler * opr-crawler * org\_bot * Perplexity-User * PerplexityBot * petalbot * pipericbot * qlyzebot * qwenbot * reflectionbot * roasanbot * rootcrawl-crawler * samparkbot * sebot * semrushbot * seojuice-searchbot * shapbot * siteinfo-bot * throxyresearchbot * triumviryresearchbot * twitterbot * uitresearchbot * vexelsearchbot * webatlabot * webxtrendbot * wpbot * wpdirectorybot * xai-searchbot * yandexbot * YouBot * zoominfobot How we count deliveries and identify bots ### Website activity and prepared delivery Cumulative retained production website delivery records, including Optiview’s own website. Known tests, operator probes, registered study exclusions, QA and demonstration routes and internal generation are excluded. Prepared-response totals count successful optimized GET responses served as Markdown or prepared HTML. Ordinary website responses, errors, redirects and bot-name-only requests without prepared output do not increase that total. Not every observed bot received prepared content. General website requests and the full bot-name inventory include original-site responses, redirects and errors. Those requests demonstrate observed activity; they are not all prepared-content deliveries. ### Identity evidence Provider IP matched means the address matched the published ranges used by our verification system. Category confirmed means Cloudflare classified the request; category verification does not independently confirm the named provider. Observed-name-only labels have neither signal. The full name inventory includes bot labels observed in request metadata, including search crawlers and other automation. Repeated spellings are merged without case differences. Names are observations, not independently authenticated provider identities. A name can be present even when its request failed or received the original website. ### Consistent activity Consistent daily activity is supported by prepared deliveries to requests with verified AI-category or provider-IP evidence on every observed day. This is an observation from retained records, not a live counter or uptime promise. ### Reading the evidence Response counts include repeat requests and are not unique bots, pages or people. Known tests are excluded; unidentified operator traffic may remain. Delivery establishes content served, not complete-body receipt, indexing, training, model ingestion, answer use or citations. Operator names do not imply a partnership or endorsement. Googlebot and Bingbot are search crawlers. Their requests do not establish use in Gemini or another AI conversation. Some observed names have no recorded prepared-content deliveries; the cards identify that distinction. Individual customer identities, request paths and raw requester metadata are not published. [Download the aggregate delivery evidence (JSON)](https://optiview.ai/proof/bot-delivery/evidence.json) Real request activity and prepared content served. Indexing, model ingestion and citations are separate outcomes. ## Additional source context Make your brand easier for AI to understand. ### Product [Website Preview](https://optiview.ai/demo/) · [Pricing](https://optiview.ai/pricing/) · [How it works](https://optiview.ai/how-it-works/) · [How it connects](https://optiview.ai/web-teams/) ### Resources [Articles & guides](https://optiview.ai/resources/) · [Get started](https://optiview.ai/onboarding/) · [Setup guide](https://optiview.ai/docs/setup/) · [FAQ](https://optiview.ai/faq/) · [Evidence & methodology](https://optiview.ai/proof/) · [For platforms & API access](https://optiview.ai/docs/) ### Company [For agencies](https://optiview.ai/partners/) · [Contact us](https://optiview.ai/#beta) · [Sign in](https://app.optiview.ai/) · [Privacy](https://optiview.ai/privacy/) ## Additional page links - [About meta-externalagent (opens in a new tab)](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) — meta-externalagent - [About Amazonbot (opens in a new tab)](https://developer.amazon.com/amazonbot) — Amazonbot - [About PerplexityBot (opens in a new tab)](https://docs.perplexity.ai/docs/resources/perplexity-crawlers) — PerplexityBot - [About Claude-SearchBot (opens in a new tab)](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) — Claude-SearchBot - [About ClaudeBot (opens in a new tab)](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) — ClaudeBot - [About ChatGPT-User (opens in a new tab)](https://developers.openai.com/api/docs/bots) — ChatGPT-User - [About DuckAssistBot (opens in a new tab)](https://duckduckgo.com/duckduckgo-help-pages/results/duckassistbot) — DuckAssistBot - [About GPTBot (opens in a new tab)](https://developers.openai.com/api/docs/bots) — GPTBot - [About Applebot (opens in a new tab)](https://support.apple.com/en-us/119829) — Applebot - [About OAI-SearchBot (opens in a new tab)](https://developers.openai.com/api/docs/bots) — OAI-SearchBot - [About Claude-User (opens in a new tab)](https://privacy.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler) — Claude-User - [About Perplexity-User (opens in a new tab)](https://docs.perplexity.ai/docs/resources/perplexity-crawlers) — Perplexity-User - [About meta-externalfetcher (opens in a new tab)](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) — meta-externalfetcher - [About Google-CloudVertexBot (opens in a new tab)](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) — Google-CloudVertexBot - [About Googlebot (opens in a new tab)](https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers) — Googlebot - [About GoogleOther (opens in a new tab)](https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers) — GoogleOther - [About bingbot (opens in a new tab)](https://developers.cloudflare.com/ai-crawl-control/reference/bots/) — bingbot - [About CCBot (opens in a new tab)](https://commoncrawl.org/ccbot) — CCBot --- # Published brand narrative # Site-wide brand information ## Who we are ### Optiview — the publishing layer for AI Optiview turns complex website content into clean, structured Markdown and combines it with the brand narrative your team chooses to publish. The result is one clear source that supported AI crawlers and other systems can retrieve, alongside your existing website experience. Built for marketing and web teams, agencies, CMS and commerce platforms, and AEO/GEO providers, Optiview helps turn recommendations into published source improvements. It complements AI visibility tools by giving teams a way to act on their findings. Applies to: Optiview ## Why we matter ### Your website\. Your narrative\. One clear source for AI\. A website can deliver a beautiful human experience while leaving important business facts scattered across pages, buried in markup, or dependent on JavaScript. AI systems may retrieve only part of that information. Optiview gives businesses control over the source they make available: readable website content plus context only their team can provide. Your web team connects delivery; your content team writes, reviews and publishes the narrative. Our focus is publishing useful, accurate brand information. Each AI service decides what it retrieves, uses and cites. Optiview does not control an LLM’s answers or guarantee placement in them. ## What we offer ### Clean website content, owned narrative and controlled delivery - **Website parser:** Converts accessible public page content into readable Markdown, retaining supported headings, text, tables and links. The parser does not use an LLM to invent facts or rewrite the page. Browser rendering is available for supported JavaScript-dependent content. - **Narrative Graph:** Organizes team-authored content into Identity, Story, Offerings, Trust, Knowledge and Actions. Add multiple items, edit them and review the graph before publishing. - **Site and page scope:** Apply an addition to the entire website or one exact page, so the published context accompanies the appropriate content. - **Publishing and delivery:** Combine eligible parsed page content with published, path-matched narrative additions through a supported website connection. Draft edits remain separate from the active publication. - **Delivery reporting:** Inspect recorded Markdown responses, requester claims, verification evidence and published versions. Delivery records show what was served, not whether it appeared in an AI answer. - **Public demo:** Try a public URL, inspect the parsed source and explore AI-drafted narrative examples. Demo results do not connect a website or publish content. ## Why trust us ### Inspect the source and understand the evidence Optiview makes the publishing process inspectable. Teams can review parser output and a combined published-output preview, check source information, and inspect publication history before deciding what to make available. The public demo provides a way to evaluate output using a real public page. Source access and extraction limitations remain visible; blocked or inaccessible content may be missing. A preview is not proof that a live request received the same output. Delivery reporting distinguishes claimed requester names from available verification evidence. A verified AI category does not by itself verify a named provider. A served Markdown response is evidence of delivery, not proof of ingestion, recommendation, citation or revenue. Learn more about the evidence and its limits at https://optiview.ai/proof/. ## Helpful knowledge ### Common questions about using Optiview **Does Optiview replace our website or CMS?** No. Optiview adds a publishing layer. It does not rewrite your CMS pages; normal HTML requests retain the website experience through supported connections. **What can we add to the Narrative Graph?** Brand descriptions, mission and audience context, products and services, supporting facts, FAQs, policies and next steps. Choose whether each item applies across the website or to one exact page. **Does saving publish our changes?** No. Saved edits remain in the draft. An authorized user publishes the graph to activate the reviewed content. Later edits stay separate until published again. **Does every AI system receive Markdown?** No. Delivery depends on the connection, request and source eligibility. Some systems explicitly request Markdown; supported delivery paths can also select it for verified AI categories. Other tools retrieve HTML and extract text themselves. **Is Optiview an AI visibility monitoring tool?** No. Optiview helps you improve and publish the source. Keep using your visibility tools to study AI answers; use Optiview’s delivery evidence to understand recorded serving activity. **Does the demo save a graph to our workspace?** No. The public demo is a separate preview. Its AI-generated proposals require review and are not automatically transferred to a workspace or published. **Do we need a developer?** Your web team may need to help establish and verify the connection. Once connected, authorized content users can manage narrative updates in the admin. **Will Optiview guarantee citations?** No. You control the source you publish. AI systems decide what to retrieve, use and cite. ## What to do next ### Try your website, then plan your connection - [Try your website](https://optiview.ai/demo/) to inspect a public page’s parsed content and explore proposed narrative additions. - [See how Optiview works](https://optiview.ai/how-it-works/) to understand the parser, Narrative Graph and combined publishing workflow. - [Explore who it’s for](https://optiview.ai/partners/) if you are an agency, CMS or commerce platform, or AEO/GEO provider bringing recommendations to customers. - [Review connection options](https://optiview.ai/web-teams/) and the [setup guide](https://optiview.ai/docs/setup/) with your web team. - Contact [hello@optiview.ai](mailto:hello@optiview.ai) to discuss access, your website and the right setup. We provide assisted workspace setup. Already connected? Review your Narrative Graph, publish when ready, inspect the combined output in the Markdown tab, and check recorded delivery in Activity. [Inspect unbranded AI conversation examples](https://optiview.ai/proof/#live-evidence), including exact prompts, answer excerpts, publication versions and matching response-body hashes. The examples are selected first-party observations, with unbranded discovery and directed URL tests labeled separately; they do not establish increased recommendation frequency.