---
optiview_delivery: edge_optimized
content_type: semantic_markdown
source_url: https://optiview.ai/resources/ai-reading-budgets/
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:33
narrative_versions:
  - 33
narrative_version: 33
---

# The AI reading budget: why less website noise matters

Your page competes with instructions, questions and other sources. Here’s why Optiview measures the text AI tools can receive—and why 10,000 tokens is a useful test window.

Research & analysis

By OptiviewFor business owners, marketers and web teamsAbout 6 minPublished November 4, 2026Sources checked November 4, 2026

- [Explore the framework](https://optiview.ai/resources/ai-reading-budgets/#interactive-framework)
- [Download guide (.md)](https://optiview.ai/resources/ai-reading-budgets.md)

ONE SHARED CONTEXT

Instructions & conversation

The customer’s question

Other retrieved sources

Your business information**Make this space count.**

Illustrative assembly · allocations vary by workflow

A shared context · illustrative assembly

An interactive overview

## Explore the framework.

Select an item to unpack the details, then take the framework with you.

4 items

01Source composition 02Extraction comparison 03Reading budget 04Findings and fixes

Item 1 of 4

### Source composition

**Business question**
How much machinery surrounds the message?

**What to do next**
Separate code and layout from readable content; inspect the actual capture.

A few terms in plain English

**Markdown**
A simple text format that keeps headings, lists and links.

**HTML**
The code used to describe a webpage.

**retrieval**
When a tool fetches information from a page or another source.

**Narrative Graph**
Optiview’s name for the business information your team organizes, reviews and approves.

Your website has a message. It also has menus, responsive layouts, scripts, repeated content and interface labels. Customers experience the design. A retrieval workflow may receive a very different representation: text extracted from the page, or selected passages from that text.

The business question is simple: how much of the material travelling with your message helps explain your business? Optiview’s AI Audit makes that question inspectable. It captures a page, compares three representations and lets you explore their size at the same reading allowance.

## Your website has to share the room.

A model’s maximum context window is the capacity of a request, not an allowance reserved for one webpage. Instructions, the question, conversation history and other retrieved material can all occupy that context. A system then chooses how much source material to include.

[Anthropic’s context-engineering guidance](https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents) treats context as a finite resource and recommends curating useful information. Our inference for website owners: give retrieval workflows a well-organized source that carries the business message without unnecessary interface baggage.

Think of a shared workspace. A larger room is useful, but it does not make every document in it relevant. A concise source can still be incomplete; a long source can still be necessary. The goal is useful information with enough supporting context.

## Why the audit starts at 10,000 tokens.

There is a real published recommendation behind this number. [Scrunch’s Site Audit documentation](https://helpcenter.scrunchai.com/en/articles/11833527-understanding-the-site-audit-tool-in-scrunch), dated July 23, 2025, recommends keeping primary content within approximately 10,000 tokens. This is vendor guidance for page content, not a universal model limit.

Optiview uses 10,000 tokens as a starting scenario. You can change it to 2,000, 5,000, 20,000 or a custom allowance. We hold the selected allowance constant across the captured HTML, standard extraction and Optiview’s prepared text. That makes their relative burden visible without pretending to know the budget of a particular assistant.

Three numbers need different labels: model capacity, the retrieval workflow’s allowance and the size of your page representation. The audit compares the third against a selected test allowance. Its opening-window preview shows the beginning of each representation. A query-based retriever can select a relevant passage from later in the page.

## Real retrieval systems already manage what gets through.

[OpenAI’s web-search documentation](https://developers.openai.com/api/docs/guides/tools-web-search) provides low, medium and high search-context settings, rather than one fixed page allowance. It also documents a returned-content budget control for supported reasoning searches. Those controls serve different purposes; neither establishes a universal 10,000-token page cap.

[OpenAI’s file-search guidance](https://developers.openai.com/api/docs/guides/tools-file-search) says limiting retrieved results can reduce token usage and latency, with a possible answer-quality trade-off. Reducing material is an engineering decision that needs relevance checks.

[Exa’s Contents documentation](https://exa.ai/docs/contents/quickstart) recommends relevant highlights for agents and factual lookups, while offering full text when broader context matters. It also supports character limits for full text. Character counts and token counts are different units. The important pattern is selection: a tool can retrieve useful passages without reading the page as one uninterrupted transcript.

[Cohere’s reranking guidance](https://docs.cohere.com/v2/docs/reranking-best-practices) documents different chunk sizes for different model generations: 4,093 tokens for v3.x and 32,764 for v4.0 before accounting for the query. Its separate 10,000-document/chunk constraint is not a 10,000-token website limit. Reranking is a configurable stage, not a universal cliff at one page size.

## Fewer tokens is only half the job.

A product claim without its qualification can become misleading. A price without the billing period is incomplete. A useful link may provide the evidence for a promise. Preparing content well means preserving those relationships while removing the material that adds no useful explanation.

[Anthropic’s Contextual Retrieval research](https://www.anthropic.com/news/contextual-retrieval) explains how isolated chunks can lose the context needed to identify and use a passage. Its retrieval experiments also illustrate the trade-off between passing more material and maintaining relevance. This is why the structure around an answer matters alongside its size.

The 2024 paper [Lost in the Middle](https://aclanthology.org/2024.tacl-1.9/) tested multi-document question answering and key-value retrieval. Changing where relevant information appeared affected performance in the models tested, with middle positions often performing worse. That is evidence to test placement and retrieval, not a claim that all current systems forget information after a fixed number of tokens.

For a business, the practical aim is clear: keep product identity, evidence, eligibility and important conditions easy to find together. Cutting a necessary disclosure to hit a smaller count defeats that purpose.

## One Apple page. A concrete comparison.

The retained Apple.com capture used in Optiview’s public example produced 8,380 tokens of whole-body standard extraction and 7,411 tokens of Optiview-prepared Markdown. The prepared version used 969 fewer tokens: an 11.6% reduction against the standard extraction.

At a 10,000-token page allowance, that leaves 1,620 tokens of room for the standard version and 2,589 for the prepared version. At 5,000, both full text versions exceed the allowance. The relative saving is still 969 tokens, but neither full version fits. Changing the window makes the difference between a measured reduction and a fit obvious.

The same example includes a captured HTML sample of 78,601 tokens. That clipped sample describes the code representation we received; it does not establish full-page HTML weight. It also does not mean an AI provider was charged for processing the HTML. The two text counts are the appropriate comparison for extracted-text reduction.

[Cloudflare’s Markdown for Agents article](https://blog.cloudflare.com/markdown-for-agents/) describes Markdown delivery as a way to reduce representation size. We keep that HTML-versus-text comparison separate from the standard-versus-prepared text comparison. The Apple example is our public website example throughout this audit.

## What the audit helps your team decide.

| Inspect | Business question | What to do next |
| --- | --- | --- |
| Source composition | How much machinery surrounds the message? | Separate code and layout from readable content; inspect the actual capture. |
| Extraction comparison | What does each text version retain? | Check products, useful links, evidence, qualifications and disclosures. |
| Reading budget | Does the full representation fit this allowance? | Change the allowance; inspect overflow and the opening-window text. |
| Findings and fixes | Which specific clutter can the team address? | Review repeated passages, link-only lists and leaked internal labels. |

The report’s room-left figures subtract the full text count from the selected page allowance. A negative result is shown as tokens over budget. The source counts remain fixed when you change the scenario. Legal text and structured data are not automatically classified as waste.

Its optional cost calculator applies your chosen input price per million tokens to the selected representation and number of reads. At that fixed rate, cost changes linearly with input tokens. It excludes output, caching, search and rendering costs. The cited sources do not establish an exponential cost jump or a universal five-second triage rule at 10,000 tokens.

[Run your free AI Audit](https://optiview.ai/ai-audit/) to inspect your own page, or [open the complete measurement method](https://optiview.ai/ai-audit-methodology.html).

## From a lighter source to a clearer business.

The audit exposes a practical problem: the representation a workflow receives can carry unnecessary material, while the information that explains the business may be scattered or incomplete. Optiview turns that observation into a publishing workflow.

Optiview prepares website content in a readable structure. Publish brings that prepared content together with business information your team reviews and approves, then makes the published source available through supported delivery. Your normal website stays in place.

The Narrative Graph addresses the information a parser cannot invent: who a service is for, what a product includes, where an offer applies, and which conditions customers need to understand. Those details may add tokens. They earn their place by explaining the business.

Our design conclusion is to optimize for useful meaning per unit of context, rather than the shortest possible document. Remove unnecessary scaffolding. Preserve evidence and conditions. Add the approved facts that are missing. Then publish a source your team can inspect and maintain.

[Explore Optiview Publish](https://optiview.ai/products/publish/) or [see your site through AI’s eyes](https://optiview.ai/ai-audit/).

Go to the source

## Keep exploring.

References linked in this guide. Open the original source for its full scope and latest details.

[anthropic.com**Anthropic’s context-engineering guidance**Opens in a new tab](https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents) [helpcenter.scrunchai.com**Scrunch’s Site Audit documentation**Opens in a new tab](https://helpcenter.scrunchai.com/en/articles/11833527-understanding-the-site-audit-tool-in-scrunch) [developers.openai.com**OpenAI’s web-search documentation**Opens in a new tab](https://developers.openai.com/api/docs/guides/tools-web-search) [developers.openai.com**OpenAI’s file-search guidance**Opens in a new tab](https://developers.openai.com/api/docs/guides/tools-file-search) [exa.ai**Exa’s Contents documentation**Opens in a new tab](https://exa.ai/docs/contents/quickstart) [docs.cohere.com**Cohere’s reranking guidance**Opens in a new tab](https://docs.cohere.com/v2/docs/reranking-best-practices) [anthropic.com**Anthropic’s Contextual Retrieval research**Opens in a new tab](https://www.anthropic.com/news/contextual-retrieval) [aclanthology.org**Lost in the Middle**Opens in a new tab](https://aclanthology.org/2024.tacl-1.9/) [blog.cloudflare.com**Cloudflare’s Markdown for Agents article**Opens in a new tab](https://blog.cloudflare.com/markdown-for-agents/)

Keep going

## A useful next read.

Research notes

### [Why publishing a page doesn’t mean AI has read it](https://optiview.ai/resources/what-openai-browser-code-taught-us/)

Developer guide

### [Fetch a public webpage as Markdown with Retrieve](https://optiview.ai/resources/webpage-to-markdown-for-rag/)

Strategy and architecture

### [A simpler page format cannot fill in missing business information](https://optiview.ai/resources/markdown-conversion-vs-approved-brand-content/)

[Back to the resource library](https://optiview.ai/resources/)

## Additional source context

Make your message easier to fit, retrieve and interpret. Remove avoidable noise while preserving the facts, links and conditions that give your business information meaning.

An overview of the table in this guide. The full reference and its qualifications remain below.

Make your business clearer to AI.

### Explore

[Publish](https://optiview.ai/products/publish/) · [Parse](https://optiview.ai/products/parse/) · [Retrieve](https://optiview.ai/products/retrieve/) · [Pricing](https://optiview.ai/pricing/)

### Learn

[Guides](https://optiview.ai/resources/) · [Developer docs](https://optiview.ai/developers/) · [Research & thinking](https://optiview.ai/research/) · [Experiments & methods](https://optiview.ai/proof/) · [AI Audit](https://optiview.ai/ai-audit/) · [AI agent directory](https://optiview.ai/agents/) · [FAQ](https://optiview.ai/faq/)

### Company

[About Optiview](https://optiview.ai/about/) · [Agencies](https://optiview.ai/partners/) · [Contact](https://optiview.ai/#contact) · [Privacy](https://optiview.ai/privacy/) · [Billing terms](https://optiview.ai/billing-policy/)

## Additional page links

- [More in this collection](https://optiview.ai/resources/topics/ai-readable-web/)
- [Research notes · Why publishing a page doesn’t mean AI has read it](https://optiview.ai/resources/what-openai-browser-code-taught-us/)
- [Developer guide · Fetch a public webpage as Markdown with Retrieve](https://optiview.ai/resources/webpage-to-markdown-for-rag/)
- [Strategy and architecture · A simpler page format cannot fill in missing business information](https://optiview.ai/resources/markdown-conversion-vs-approved-brand-content/)


---

# Published brand narrative

# Site-wide brand information

## Who we are

### Optiview — Parse, Retrieve and Publish for AI

Optiview helps businesses and developers make website information easier for AI to use. Parse converts supplied HTML into clean Markdown and source metadata. Retrieve fetches an eligible public HTTPS webpage and prepares its content. Publish helps a verified business prepare its own website content and make approved Narrative Graph context available through a supported connection.

Parse and Retrieve use application keys and a prepaid developer wallet. Publish has separate website setup, ownership checks and a subscription. Optiview does not control whether an outside AI service discovers, uses or cites the resulting content.

Applies to: Optiview

## Why we matter

### A clearer source, from website content to business context

Webpages mix useful information with navigation, layout and application code. Businesses may also have important qualifications, answers and context that are not clearly published.

Optiview gives developers a way to prepare source content for their applications, and gives businesses a way to publish the context their teams approve. Your application chooses how to use Parse and Retrieve results. Your business controls what it publishes through Publish.

Clearer publishing makes information available; it does not determine an AI system’s answer. Each outside service chooses its own retrieval, recommendations and citations.

## What we offer

### Three products, with separate jobs and clear pricing

- **Optiview Parse:** Supply complete HTML and a source URL. Receive clean Markdown, source and output hashes, parser details and processing warnings. Parse does not fetch the URL or execute JavaScript. A successful page uses 1 credit.
- **Optiview Retrieve:** Supply one eligible public HTTPS URL. Optiview checks supported access and publisher policies, fetches anonymous static HTML and parses it. It does not search the web, execute JavaScript, bypass a login or access private networks. A successful page uses 5 credits total, including parsing.
- **Developer credits:** USD $20 buys 20,000 credits shared by Parse and Retrieve. There is no developer subscription, automatic recharge or scheduled credit expiration. Failed jobs consume no credits. Applicable taxes and the total appear at checkout. Source size, quotas and retention limits apply.
- **Optiview Publish:** USD $99 per website per month plus applicable tax. Connect a verified website, review the prepared source and optionally maintain approved Narrative Graph context in Identity, Story, Offerings, Trust, Knowledge and Actions. Choose site-wide or exact-page scope. Saving drafts does not publish them. Eligible AI tools can receive prepared Markdown or HTML through supported delivery, while ordinary visitors keep the normal website.

Developer credits do not purchase Publish subscriptions. Parse and Retrieve do not add a customer’s private Narrative Graph to arbitrary third-party pages. See https://optiview.ai/pricing/ and https://optiview.ai/developers/ for the current offer and limits.

## Why trust us

### Inspect the output and understand what the evidence establishes

Developers can inspect the prepared Markdown, job status, source metadata, processing warnings and credit ledger. Retrieve records the public source response it observed; Parse identifies supplied HTML rather than claiming to verify the current origin.

Publish administrators can review parser output, the combined published-output preview, source checks and publication history. A preview is separate from a recorded live delivery. Saved edits remain separate from the active publication until explicitly published.

Our experiment pages contain selected first-party AI conversations and source comparisons. Delivery records establish what was served, not downstream ingestion, recommendations, citations or revenue. Observed language matches do not establish causation or a consistent visibility improvement. Explore the evidence and limitations at https://optiview.ai/proof/.

## Helpful knowledge

### Common questions about the three products

**Which product should I use?**
Use Parse when you already have HTML. Use Retrieve when you need Optiview to fetch one eligible public page first. Use Publish when your business wants to prepare its verified website and publish approved context.

**Do Parse and Retrieve require website ownership or DNS changes?**
No website connection is needed for these developer products. You remain responsible for source rights and permitted downstream use. Public availability does not grant a reuse license.

**How do developer accounts and keys work?**
Create a developer workspace and save the personal sign-in key and offline recovery code. Create scoped service keys for server-side applications. Email is an unverified contact identifier, not account-ownership proof or an email-only recovery method. Keep all credentials private.

**Can my agent use Optiview?**
The authenticated developer API returns prepared content for your application or model. A downloadable JavaScript client and local MCP adapter are available. Your host must authorize sources and any credit expenditure. Compatibility does not establish adoption or a partnership with an AI provider.

**Does the API include my private Narrative Graph?**
Parse and Retrieve are source-only. Verified website publishing and approved publisher context belong to the separate Publish workflow.

**Does Publish replace our website or CMS?**
No. Your website stays in its existing system. Your web team establishes and tests a supported connection. Your team reviews and explicitly publishes any Narrative Graph additions. Ordinary visitors retain the normal website experience.

**Does every AI tool get the prepared version?**
No. Delivery depends on the website connection, source eligibility and the request. Some tools explicitly request Markdown; supported AI requests may also receive prepared Markdown or HTML through eligible delivery paths. Traditional search crawlers retain the original website response. A claimed tool name does not prove the requester’s identity.

**Does the demo publish or save to our workspace?**
No. The public demo is a separate preview. Its AI-generated examples require review and do not transfer automatically into an approved publication.

**Will Optiview guarantee citations or recommendations?**
No. Optiview prepares and publishes sources. Outside AI services choose what to retrieve, use and cite.

## What to do next

### Choose your starting point

- [Compare Parse, Retrieve and Publish](https://optiview.ai/products/) to choose the right workflow.
- [Create a developer workspace](https://app.optiview.ai/developers) for self-service Parse and Retrieve, then add credits and create an application key.
- [Read the API quickstart and agent-tool setup](https://optiview.ai/developers/) for request examples, supported sources, errors, quotas and downloads.
- [Review pricing and terms](https://optiview.ai/pricing/) for prepaid credits and the separate Publish plan.
- [Preview your website](https://optiview.ai/demo/) to inspect prepared content and explore unapproved narrative examples without connecting or publishing.
- [Plan Publish setup](https://optiview.ai/onboarding/) with your web team. Publishing requires ownership checks, review and a supported connection; contact [hello@optiview.ai](mailto:hello@optiview.ai) for assisted setup.

Already using Publish? Review your Narrative Graph, publish approved changes explicitly, inspect the combined source in Markdown and check recorded delivery in Activity. Keep your existing visibility tools to study downstream AI answers.