---
title: "Every document opened, with its origin."
seo_title: "See what is inside your documents | SazAI"
description: "Every document becomes pages in your vault, and the pages become a network you can search and trace."
tags: [marketing, dimensao/veja]
icon: eye
---

> Someone joins your team and finds twelve presentations. A client sends a two-hundred-page manual due tomorrow. SazAI Corpus reads all of it once and hands back every document opened, in [four formats](documents).

<div class="chap">

<div class="ico-col"><sz-icon style="color:var(--verb-see)" name="file" size="36"></sz-icon></div>


<h2 id="in-and-out">What comes in, and what comes out</h2>

<h3 class="sub">One file goes in, four formats come out</h3>


The presentation becomes one section per slide; the document, one per section; the
spreadsheet, one per sheet; the PDF, one per page. Each section carries the image, the
text and, when they exist, the speaker notes and the comments. The twelve presentations
become twelve places you browse; the manual becomes two hundred sections you move through
instead of reading.

<figure class="viz duo strip" role="group" aria-label="One file goes in and becomes four kinds of page: slide, section, sheet and page">
<div class="side"><svg viewBox="0 0 24 24" fill="none" stroke="var(--gray)" stroke-width=".75" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true" style="width:56px;height:56px" data-icon="file"><path d="M13 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V9z"/><polyline points="13 2 13 9 20 9"/></svg><b>one file</b>goes in once</div>
<div class="mid"><svg viewBox="0 0 120 72" aria-hidden="true"><line x1="10" y1="36" x2="60" y2="36" stroke="var(--verb-see)" stroke-width="1.5"/><path d="M54 31l6 5-6 5" fill="none" stroke="var(--verb-see)" stroke-width="1.5"/><line x1="60" y1="36" x2="100" y2="14" stroke="var(--gray)" stroke-width="1.5"/><line x1="60" y1="36" x2="100" y2="29" stroke="var(--gray)" stroke-width="1.5"/><line x1="60" y1="36" x2="100" y2="43" stroke="var(--gray)" stroke-width="1.5"/><line x1="60" y1="36" x2="100" y2="58" stroke="var(--gray)" stroke-width="1.5"/><circle cx="60" cy="36" r="3" fill="var(--verb-see)"/></svg><b>and becomes pages</b>in your format's unit</div>
<div class="side out4">
<span><svg viewBox="0 0 24 24" fill="none" stroke="var(--verb-see)" stroke-width=".75" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true" style="width:44px;height:44px" data-icon="file-pptx"><rect x="3" y="4" width="18" height="13" rx="1"/><line x1="9" y1="21" x2="15" y2="21"/><line x1="12" y1="17" x2="12" y2="21"/></svg><b>slide</b>presentation</span>
<span><svg viewBox="0 0 24 24" fill="none" stroke="var(--verb-see)" stroke-width=".75" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true" style="width:44px;height:44px" data-icon="file-docx"><path d="M13 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V9z"/><polyline points="13 2 13 9 20 9"/><line x1="8" y1="13" x2="16" y2="13"/><line x1="8" y1="17" x2="14" y2="17"/></svg><b>section</b>document</span>
<span><svg viewBox="0 0 24 24" fill="none" stroke="var(--verb-see)" stroke-width=".75" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true" style="width:44px;height:44px" data-icon="file-sheet"><rect x="3" y="3" width="18" height="18" rx="1"/><line x1="3" y1="9" x2="21" y2="9"/><line x1="3" y1="15" x2="21" y2="15"/><line x1="9" y1="3" x2="9" y2="21"/></svg><b>sheet</b>spreadsheet</span>
<span><svg viewBox="0 0 24 24" fill="none" stroke="var(--verb-see)" stroke-width=".75" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true" style="width:44px;height:44px" data-icon="file-pdf"><path d="M13 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V9z"/><polyline points="13 2 13 9 20 9"/><rect x="8" y="12" width="8" height="6" rx="1"/></svg><b>page</b>PDF</span>
</div>
</figure>

The file knows more than the presentation showed: the slide someone hid appears for you, marked as hidden.

</div>

<div class="chap">

<div class="ico-col"><sz-icon style="color:var(--verb-see)" name="search" size="36"></sz-icon></div>


<h2 id="documents">Your documents</h2>

<h3 class="sub">One reading on entry, and everything it produced gets an address</h3>

An Office file is a package, and SazAI Corpus reads the whole package, once, through an
engine that reads the format's specification and does not interpret it. The <dfn tabindex="0" data-hint="All the text of the document, each passage tied to the slide and shape it came from. Extracted once, without interpretation.">X-ray</dfn> is what
comes out of that reading: the text with its origin, the comments with their state, the
notes, the hidden items, the properties. And what the engine measured has a name; your
editions enter as complete documents, and the comparison returns a number instead of an
impression.

Every claim in the document gets an address, and the address belongs to the file, not to
our opinion about it. The whole of it, with the receipt of what changed and what it does
not prove, is in [your documents](documents).

</div>

<div class="chap">

<div class="ico-col"><sz-icon style="color:var(--verb-see)" name="compare" size="36"></sz-icon></div>


<h2 id="use-cases">Use cases</h2>

<h3 class="sub">Where this solves something, and where it does not yet</h3>

Understanding material you did not produce. Comparing versions and knowing what changed.
Translating a whole training program and being able to prove it arrived intact. Analyzing
a portfolio instead of one file at a time.

The situations did not come from market research; they came from watching the same scene
repeat across different sectors, and each one separates field testimony from the
verifiable claim. The ones we have seen up close, together with what is not yet solved,
are in [use cases](use-cases).

</div>

<div class="chap">

<div class="ico-col"><span class="ico" style="color:var(--verb-see)"><svg viewBox="0 0 48 48" aria-hidden="true"><circle cx="14" cy="24" r="4" fill="none" stroke="currentColor" stroke-width="1.5"/><circle cx="34" cy="14" r="4" fill="none" stroke="currentColor" stroke-width="1.5"/><circle cx="34" cy="34" r="4" fill="none" stroke="currentColor" stroke-width="1.5"/><line x1="18" y1="22" x2="30" y2="16" stroke="currentColor" stroke-width="1.5"/><line x1="18" y1="26" x2="30" y2="32" stroke="currentColor" stroke-width="1.5"/></svg></span></div>


<h2 id="connectome">Neural network</h2>

<h3 class="sub">Documents stop being a folder and become a diagram</h3>

Nothing sits loose in your vault. Every document is linked to its editions, every edition
to its receipt, every comparison to both sides. The <dfn tabindex="0" data-hint="The graph of your vault: every document a node, every link a synapse. It grows with every upload; the agent reads through it, node by node.">connectome</dfn> is that diagram: every
document a node, every link a synapse, and it grows with every upload without you
assembling anything.

You move through it on screen. [The agent](talk) moves through it from inside, node by
node, with the origin and the measurement already sitting beside it, which is why the
answer can come with its source. The whole network, and why its links are declared and
never guessed, is in [neural network](connectome).

</div>

The engine measures and does not opine. The one who interprets what was measured is [the agent](talk) of SazAI Corpus, reading the same vault as you, with no source beyond your own.
