Universal Corporate PDF Standards Intelligence Workbench

No PDF loaded

Click Open PDF. All indexing stays in this browser.

Waiting for PDF...

Text Layer: Not checked Standards: 0

Guided Workbench

Guide

Purpose: Convert a practical design or review task into a standards review path. Choose a workflow below or type your own task, then open the linked source pages to verify the requirements.

Open a PDF, then use a task button or describe the design task. The Workbench will create a standards review path, source cards, related references, and checklist actions.

Revision intelligence appears here after indexing.

Standards Index

Open a PDF to build a standards index.

Section Navigator

Open a PDF to build the section menu.

Smart Search

Smart results will appear here after indexing.

Assistant

Local

AI Settings Not configured

Provider

API Key

Model

Your API key is stored only in this browser's localStorage and is sent directly to the AI provider when you ask a question. Never paste keys on a shared device. The PDF text is never uploaded — only relevant excerpts are sent per question.

The assistant gives document-grounded results with supporting pages and confidence.

Design Checklist

Generate a checklist from a task, assistant question, or current standard.

Designer Notes

Notes are saved locally per PDF checksum.

Research Packet

Export

Open a PDF and use the Workbench, Assistant, Pins, Notes, or Checklist to build a project research packet.

Pinned Standards

Pin frequently used standards or pages for quick access.

History by Topic

Assistant history is saved locally per PDF checksum and grouped by topic.

Document Library

PRO

Load multiple PDFs simultaneously and switch between them instantly. Enable cross-document search to find standards across your entire library at once.

Search across all documents

No documents in library. Open a PDF to get started.

Help Center

Guide

What this viewer is for

This tool turns any local PDF with a readable text layer into a fast standards-finder workspace. It is intended for engineering standards, construction manuals, specification books, policy manuals, technical drawings, and other large reference PDFs.

Operating model: select a PDF, let the browser index it locally, then use Standards, Sections, Smart Search, Assistant, Pins, History, and the PDF viewer together.

Privacy model: the PDF is processed locally in the browser. No cloud API, batch file, PowerShell, Python, executable, or internet connection is required after extraction.

1. Quick Start

Extract the ZIP folder first. Do not run the viewer from inside the compressed ZIP preview.
Open index.html in Chrome or Edge.
Click Open PDF and select a local PDF.
Wait until the status card says the document is ready.
Use Quick Find for fast lookups, Standards for detected standard numbers, Sections for document structure, Search for ranked matches, or Assistant for question-based lookup.
When you are ready to read the PDF, click Focus PDF in the PDF toolbar to minimize the left workbench and give the document more screen space.

Important: the left panel now uses one full-height sidebar scroll. If a tab has many cards, scroll the left sidebar itself to reach all results.

2. Top Bar and Open PDF Controls

Control	Purpose	Best use
Open PDF	Selects a local PDF and starts local indexing.	Use whenever you want to load a new manual or reference document.
Hide Sidebar	Hides only the left Workbench/sidebar area and lets the PDF viewer use the extra width.	Use when you still want the header, Quick Find, and PDF toolbar available but need a wider PDF reading area.
Reload View	Re-renders the current PDF page if the canvas view appears stale.	Use after browser zoom changes or if a page preview looks incomplete.
Close PDF	Clears the currently loaded document from the active viewer session.	Use before switching to unrelated documents.
Quick Find	Fast command-style lookup across standards, sections, page headings, and topics.	Use first when you know a code, section number, or topic keyword.
Focus PDF	Minimizes the left workbench/sidebar and hides Quick Find results so the PDF viewer expands across the screen.	Use after locating the correct standard or page and you need maximum viewing space.

2A. Hover Instruction Cards

The viewer now includes contextual instruction cards. Hover your mouse over major buttons, tabs, filters, result cards, status badges, workflow buttons, checklist items, notes, history cards, and PDF controls to see quick operating guidance.

Mouse users: hover over a clickable item or information card.
Keyboard users: tab to a button, input, or control to show the same instruction card.
Efficiency tip: use hover cards to learn what each control does without leaving the active workflow.

These cards are local UI help only. They do not call the internet, change the PDF, or upload document content.

3. Status Card, Progress, and Quality Badges

The status card tells you what the viewer is doing while it reads the PDF. Large manuals may take time because every page is processed locally.

Progress bar: shows indexing progress.
Text Layer badge: estimates whether the PDF has usable searchable text.
Standards badge: shows how many standard/drawing-style identifiers were detected.

Text layer status	Meaning	Recommended action
Good	Most pages contain extractable text.	Search and Assistant should work well.
Partial	Some pages have text and some may be scanned or graphic-heavy.	Verify results manually; OCR may improve accuracy.
Weak	The PDF is likely image-only or poorly OCR’d.	OCR the PDF before relying on Search or Assistant.

4. Quick Find Command Bar

Quick Find is the fastest way to locate known standards, section numbers, common topics, drawing numbers, page headings, or short phrases.

AB-1200UG3002102section 9any terma headinga defined word

Best practice: start short. Search a single distinctive word before trying a long sentence. Use the Assistant for longer questions.

5. Standards Tab

The Standards tab lists detected standard, drawing, or page-code identifiers. It works best with documents that contain labels such as OH102, UG3002, STD-101, or similar technical identifiers.

Filter box: narrows the Standards list by code, title, page text, or section.
Open / card click: jumps to the detected page.
Pin: saves frequently used standards to the Pinned tab.
Related: helps find nearby or similar pages after opening a standard.

If the Standards tab finds few items, the PDF may not use recognizable standard numbers. Use the Sections or Search tabs instead.

6. Sections Tab

The Sections tab provides document navigation. It uses PDF bookmarks/outlines when available. If the PDF has no bookmarks, it creates fallback page groups from visible text headings and page structure.

Collapsible groups: browse by detected section or heading group.
Section filter: search within section names, page headings, or grouped pages.
Page jump: selecting a section card renders the matching PDF page.

Best use: use Sections when you want to browse the manual structure instead of searching individual words.

7. Search Tab / Smart Search

Smart Search returns ranked cards instead of a flat text list. The ranking prioritizes exact standard numbers, headings, section matches, and then body text.

How to use it:

Enter a short topic or code.
Click Search PDF.
Review the highest-scoring cards first.
Use filters only after the first broad search if there are too many results.

Filter	What it does	Example use
Numbered items only	Focuses on pages carrying a detected identifier (standard, clause, or drawing number).	When you want the controlling item, not every passing mention.
Requirements	Prioritizes shall / must / required / prohibited language.	Finding binding rules.
Findings & recommendations	Prioritizes observations, findings, deficiencies, and recommendations.	Working through a report, audit, survey, or inspection.
Notes & cautions	Prioritizes notes, cautions, warnings, and callouts.	Finding exceptions and advisories.
Procedures	Looks for steps, sequences, methods, and instructions.	Finding how something is done.
Tables & figures	Focuses on tabulated data, figures, dimensions, ratings, and quantities.	Looking up a specific value.
Definitions	Prioritizes glossary entries, defined terms, and abbreviations.	Confirming how this document uses a term.
Cross-references	Prioritizes pointers to other sections, appendices, or documents.	Tracing dependencies.
Revisions	Looks for revision history, changes, and superseded content.	Checking what changed.

If result cards are longer than the visible panel, scroll the entire left sidebar to see the remaining cards.

8. Assistant Tab

The Assistant is a local document-grounded lookup helper. It does not call a cloud AI service. It searches the indexed PDF and builds an answer from the strongest matching pages.

Strong assistant questions:

What does this document require for this task?
Where are the inspection requirements discussed?
Which pages list the materials or inputs for this topic?
Show the best pages for pedestrian accessibility.
Which section covers this topic?

Output format: the Assistant provides a summary, source result cards, supporting snippets, and confidence guidance.

Accuracy rule: always open and verify the supporting source pages before using the result for design, construction, safety, or compliance decisions.

9. Pinned Tab

The Pinned tab stores pages or standards you use often. Pins are saved locally in your browser and separated by PDF checksum, so each PDF has its own pin set.

Pin standards: use this for pages you repeatedly reference.
Open pinned item: jumps back to the saved page.
Clear Pins: removes the saved pins for the current PDF.

Recommended workflow: during a project, pin every standard/page you expect to revisit. Use Pinned as your project reference shortlist.

10. History Tab

The History tab stores Assistant questions and responses locally by PDF checksum and groups them by topic. This creates a lightweight research trail for each document.

Topic grouping: similar questions are grouped together.
Export: downloads a JSON file of the current PDF’s assistant history.
Clear: deletes local assistant history for the current PDF.

History can disappear if browser storage is cleared, a different browser is used, private/incognito mode is used, or a revised PDF has a different checksum.

11. Help Tab

The Help tab is the built-in training guide. It explains the viewer controls, search strategy, assistant behavior, security model, and troubleshooting steps.

For teams, use this Help tab as the first-pass onboarding guide before distributing the viewer to other users.

12. PDF Viewer Area

The right side includes two PDF viewing modes. Enhanced Workbench uses PDF.js canvas rendering with an annotation/link layer, so internal PDF links that are detected as PDF annotations can be clicked while the Workbench remains fully active. Native PDF Links uses the browser’s built-in PDF viewer so users can use the PDF’s original navigation/bookmark/link behavior when the browser allows blob PDF viewing.

Use Enhanced when you want the strongest integration with Standards, Sections, Search, Assistant, Workbench, pins, notes, and controlled page jumps. Use Native Links when you want to browse the PDF exactly like a normal PDF with its own internal links.

Control	Purpose
Previous	Moves back one PDF page.
Page number box	Type a page number and press Enter to jump directly.
Next	Moves forward one PDF page.
Related standards/pages	Shows topically similar or nearby pages after opening a page or result.

Page jumping: Standards, Sections, Search, Assistant, Pins, and History all use the same page-rendering jump method.

13. Left Sidebar Scrolling

The left side is intentionally designed as one scrollable sidebar. This prevents nested card lists from collapsing into each other on smaller screens or high browser zoom levels.

Click the tab you need: Standards, Sections, Search, Assistant, Pinned, History, or Help.
Use the mouse wheel or trackpad while the pointer is over the left sidebar.
The tab row remains available while you move through long result sets.
The PDF viewer on the right remains independent from the left sidebar.

If you cannot see more cards, scroll the left sidebar itself instead of trying to scroll inside a single card.

14. Recommended Workflows

Known standard number: type the code in Quick Find → open the top result → pin it if used often.

Unknown topic: use Smart Search with a broad term → review best cards → add filters only if needed.

Question-based lookup: ask Assistant → review confidence and source cards → open the supporting pages.

Browsing a large manual: use Sections → filter section names → open likely section/page cards.

Project reference set: search once → pin all relevant pages → work from the Pinned tab.

Multiple PDFs: load one PDF at a time. Pins and history stay separated per document checksum.

15. Guided Workbench Tab

The Workbench converts a practical task into a standards review path. Use it when you do not know the exact page or standard number.

Task buttons: generated from the open PDF — its most frequent subjects, plus reviews for the kinds of content it actually contains. They change with every document.
Build Review Path / Find Standards: type a design task and the app ranks likely standards/pages, creates recommended review steps, shows source cards, and prepares a checklist.
Standard Summary: summarizes the current page or selected workflow using extracted PDF text only.
Revision Intelligence: highlights revision/change pages detected in the PDF.

16. Design Checklist Tab

The Checklist tab turns search results into a task checklist with linked source pages. It is designed for design review, QA/QC, and field verification preparation.

Generate a checklist from a topic, Workbench task, or Assistant result.
Click source buttons beside checklist items to open supporting pages.
Checked items are saved locally per PDF checksum.
Use Reset Checklist to clear the current checklist for the open PDF.

17. Notes and Research Packet Tabs

Designer Notes lets users save local notes per page or standard. Use this for internal reminders, field observations, review comments, or project-specific interpretation notes.

Research Packet compiles document metadata, current checklist status, pinned pages, notes, assistant history, and recent source pages into one exportable text file.

These tools do not change the PDF. They save local browser data tied to the PDF checksum.

18. Annotation Tools

The annotation bar sits directly above the PDF page in Enhanced Workbench mode. It appears once a PDF is open.

Hl / Un / St — highlight, underline, or strike through. Pick the tool, then select text on the page and the mark is applied to your selection.
Draw — freehand pen. Three line widths are available: thin, medium, and thick.
Note — click anywhere on the page to drop a sticky note and type its text. Click an existing note marker to reopen and edit it.
Erase — click an annotation to remove it. Right-clicking any annotation also deletes it, whatever tool is active.
Colour swatches — set the colour used by whichever tool is active. Changing colour does not alter marks that already exist.

Save PDF rewrites the open document with every annotation embedded and downloads it as <filename>-annotated.pdf. The result is an ordinary PDF that opens correctly in Acrobat, a browser, or any other reader.

Annotations are stored in this browser, keyed to the PDF's checksum, and are restored when you reopen the same file. Your original file on disk is never modified. Clearing browser data deletes them — use Save PDF for anything you need to keep.

19. OCR for Scanned Pages

When a page has little or no extractable text, a warning bar appears above the PDF offering OCR. This is how you make a scanned document searchable.

OCR this page — reads just the page you are viewing.
OCR all pages — processes every page that looks text-poor, leaving good pages untouched.
Recovered text is merged into the index, so Search, Assistant, Standards, and Sections start working for those pages.

OCR downloads its recognition engine (Tesseract) from a public CDN the first time you use it, so this one feature needs an internet connection and must be permitted by your network policy. The PDF itself is still never uploaded — recognition runs inside your browser. OCR is slower than normal indexing and its accuracy depends on scan quality.

20. AI Assistant (optional)

By default the Assistant is a purely local search-and-ranking helper. If you want conversational answers, open AI Settings on the Assistant tab and supply your own API key.

Provider — leave as Local search only to keep everything offline, or choose Anthropic (Claude) or OpenAI (GPT).
API key — your own key. It is stored only in this browser's local storage.
Model — faster/cheaper or higher-quality options are listed per provider.
Answers stream into the Assistant tab and are shown alongside the source pages used as context.

Enabling this sends data off your machine. Relevant page excerpts — not the whole PDF — are sent directly to the provider you choose, along with your question. Do not enable it for confidential documents without clearing it with your organization first, and never paste an API key on a shared device. Use Clear Key to remove a stored key.

21. Docs — Document Library

The Docs tab keeps several PDFs open at once. Each document you open is added to the library with a thumbnail and page count.

Click any library entry to switch to it. The whole left panel — index, sections, checklist, notes, pins, history, and annotations — switches with it.
Each document keeps its own state, so switching away and back loses nothing.
Search across all documents widens Smart Search to every PDF in the library, with results labelled by source document.
Use the × on an entry to remove that document from the library.

The library lasts for the current browser session. Pins, notes, checklists, and annotations are saved per PDF checksum, so reopening a file later restores its work.

22. Troubleshooting

The PDF indexes but search results are weak

The PDF may be scanned/image-only or have poor OCR. The viewer can display it, but Search and Assistant need extractable text.

Search returns too many results

Start with fewer words, then add one filter. Exact standard numbers and short terms usually work better than long sentences.

Search returns too few results

Remove filters, try synonyms, or use a shorter term. Example: search one distinctive word instead of a long phrase.

Cards appear to stop before all results are visible

Scroll the entire left sidebar. The current layout avoids nested scrolling by allowing the whole left panel to move vertically.

Indexing is slow

Large PDFs are processed page by page locally. Keep the tab active and wait for the status card to show Ready.

History or pins disappeared

Pins and history are stored in browser local storage per PDF checksum. Clearing browser data, using another browser, private browsing, or loading a revised PDF creates a separate local set.

PDF page does not render clearly

Use Reload View or adjust browser zoom. Extremely large drawing sheets may take extra time to render.

23. Corporate and Security Notes

No batch files, PowerShell scripts, Python scripts, Java files, installers, or executables are required.
No internet connection is required for normal use. Two optional features are the exception: OCR downloads its engine from a public CDN, and the AI Assistant contacts the provider you configure. Both are off unless you turn them on.
No PDF is uploaded to a server by this app. If you enable the AI Assistant, relevant page excerpts are sent to your chosen provider with each question.
PDF content is processed locally using bundled PDF.js files.
Assistant history and pins use browser local storage.
The package is easier for corporate review because it is static HTML, CSS, JavaScript, and PDF.js assets only.

24. Limitations

The Assistant is a local search-and-ranking helper by default, not a cloud large language model. Cloud answers are available only if you supply your own API key under AI Settings.
It cannot read text that is only present as a flat image unless the PDF already has OCR text, or you run the built-in OCR on those pages.
Automatic standards detection depends on recognizable document codes or headings.
Native PDF Links mode is controlled by the browser PDF viewer. If a user clicks a link inside the native PDF pane, the viewer may not always report the new internal PDF page back to the Workbench toolbar.
Always verify final answers against the opened source pages before using them for formal decisions.

Open a PDF to begin

This viewer reads the selected file locally, builds smart navigation/search/assistant results, and displays the PDF here.

Key	Action
`←` or `↑`	Previous PDF page
`→` or `↓`	Next PDF page
`+` or `=`	Zoom in (Enhanced mode)
`−`	Zoom out (Enhanced mode)
`0`	Reset zoom to fit width
`Ctrl`+`Shift`+`F`	Toggle PDF Focus Mode (hides sidebar and Quick Find)
`Esc`	Close this help guide / dismiss tooltip

Query type	Example	What you get
Standards code	`OH102`, `UG3002`	Exact match at the top, then related pages
Section number	`section 900`, `1100`	Section group cards with page links
Topic / keyword	`two key words`	Ranked pages by relevance
Plain question	`what is the maximum spacing`	Same as keyword — top matching pages

Workflow	What it targets
Document topics	Up to six buttons named after the subjects that recur most across this PDF's own headings
Requirements	Pages stating obligations, limits, or conditions
What Changed	Revision notes, updates, and superseded content
Tables & Figures	Tabulated values, dimensions, and figures
Definitions	Defined terms, glossary, and abbreviations
Procedures	Step-by-step methods and instructions

Tab	Purpose	Best used when
Workbench	Guided task-based review paths with source cards	You have a design or construction task and need to find applicable standards
Standards	Auto-detected standard/drawing codes grouped by series	You know a code number and want to find its pages directly
Sections	Document structure using PDF bookmarks or heading detection	Browsing chapter structure or navigating large manuals
Search	Full-text ranked search with type and tag filters	Looking for a topic across the whole document with more control
Assistant	Document-grounded Q&A from indexed page text	Asking plain-language questions about specific requirements
Checklist	Task checklists with linked source pages and completion tracking	Design review, QA/QC prep, or field verification steps
Notes	Per-page or per-standard local notes saved by PDF checksum	Recording interpretation notes, field observations, or review comments
Packet	Exportable research packet combining standards, pages, and notes	Compiling a deliverable or sharing findings with a team
Pinned	Saved pages and standards for quick return navigation	Frequently referenced pages or standards you open repeatedly
History	Local record of Assistant questions and answers by PDF	Reviewing prior research or continuing a previous session

Mode	Description	Use when
Enhanced Workbench	PDF.js canvas renderer with annotation link layer, controlled page jumps, and full Workbench integration	Using source cards, page jumps, workbench workflows, or zoom/pan
Native PDF Links	Browser built-in PDF viewer — preserves the PDF's original bookmarks, TOC, and annotation links	Navigating with the PDF's own built-in table of contents or link structure