A full construction documentation set is typically 100-500 pages: architectural plans, structural drawings, MEP overlays, schedules, details, specifications. Finding a specific door tag, room number, equipment ID, or detail callout across that volume should be instant. In most PDF viewers it isn't, the search runs sequentially through every page, opening each in turn, with visible lag between hits. For coordination meetings and RFI responses where speed matters, the search experience is the actual bottleneck more often than the rendering.

Why search in large PDFs is slow in general viewers

Most PDF viewers index a document's text content on demand, meaning the first search on a freshly-opened file triggers a full text extraction across every page. For a 200-page construction set, that initial pass can take 5-15 seconds in Adobe Acrobat, Apple Preview, Foxit, and Nitro PDF. Subsequent searches are faster (the index is cached), but every fresh open repeats the cost. For architects who open dozens of project PDFs in a day, switching between disciplines, between revisions, between projects, the initial search lag becomes a meaningful productivity drain. Worse: when searches return many hits across multiple pages, navigation between hits often re-renders each page from scratch.

How Ncored handles search on multi-page sets

Open the set, press Cmd/Ctrl+F, and type. You are not waiting for the drawing to finish anything first, and the first search of the session is as quick as the tenth. Jumping between hits stays quick too, including on the busy sheets where a search usually lands. Cmd/Ctrl+F looks through the text and the annotations on every page, so a note somebody added during review is found the same way as a title block label. For an RFI meeting where you need door D-23 across the architectural and the structural sheet, that is the difference between answering in the meeting and promising to come back.

Search is ready when you are
Press Cmd/Ctrl+F and type. The first search of the session is as quick as the tenth.
Annotations are searched too
Cmd/Ctrl+F looks through the text and the markups on every page, so review notes are found like any other text.
Moving between hits stays quick
Stepping through results across different sheets does not get slower as you go.
Works on CAD-exported sets
Tested on exports from ArchiCAD, Revit, AutoCAD, Vectorworks and SketchUp Layout: labels, callouts and dimension text are all found.
Scanned sheets with an OCR layer
If the PDF already carries an OCR text layer, from ABBYY FineReader or an Acrobat OCR pass, it is searched like native text.

Try the 14-day free trial

Download Ncored

Search performance comparison

Adobe Acrobat, Foxit and Nitro PDF all appear to prepare their search on demand, so the first search on a freshly opened 200-page set is the slow one, commonly five to fifteen seconds before anything appears, and the cost returns with every fresh open. Apple Preview has the simplest search of the group and is fine on light documents. Bluebeam Revu has a well-developed search for its own workflow, and on Mac it is not an option at all since Revu for Mac reached end of life in 2023. With Ncored the first search behaves like the tenth, which matters most for the people who open dozens of sets in a day and search each one once.

Frequently Asked Questions

Can Ncored search across multiple PDFs at once?
Currently no, search is per-document. Cross-document search across an entire project archive is on the roadmap but not in the current release. For now, the workflow is: open one PDF, find what you need, open the next.
Does search work on scanned construction drawings?
Only if the scanned PDF already has an OCR text layer embedded (typical for documents OCR'd in Acrobat or ABBYY before being sent to you). Ncored doesn't run OCR itself, if the PDF is image-only with no embedded text, search won't find anything because there's nothing to index. The fix is to OCR the file first in Adobe Acrobat or a dedicated OCR tool.
Does search match partial words?
Yes. Searching 'D-2' matches D-23, D-205, all door tags starting with D-2. This is the expected behavior, architects rarely remember exact tag suffixes.
What about searching dimension values like '3.45m'?
Dimension text in CAD-exported PDFs is stored as standard text labels and is fully searchable. Searching '3.45' finds every dimension chain in the set that uses that value. Searching 'm' matches everything ending in m, usually too noisy to be useful; better to search the specific value.
What if my construction drawings are scanned without an OCR text layer?
Then they're effectively images, not searchable text, Ncored can't search what doesn't exist as text in the PDF. The fix is to run OCR first (Adobe Acrobat's OCR is the standard, ABBYY FineReader is the precision option for engineering documents), then open the OCR'd version in Ncored. Once an OCR text layer exists in the PDF, Ncored indexes it identically to native CAD-exported text.