Make PDF Searchable Online: From Upload to Saved Copy

To make a PDF searchable online, open a web OCR tool in your browser, upload a copy of the file, wait while the pages are read, and download the searchable PDF it returns. Nothing is installed. The job is finished when the copy is saved in the matter file and the upload is deleted from the website.

What happens to a PDF between the upload and the searchable download?

Between upload and download a web tool does four things: it receives your file, reads each page picture into text, builds a new PDF that carries the text with the pictures, and holds the result until you fetch it.

The transfer and the reading are the same for any online conversion, and the post on online text recognition follows a page between upload and recognized text step by step. Where the reading runs depends on the tool. A server-based tool reads the file on the provider's servers, and some tools read it inside your browser; the same post explains how to tell which you have.

Save the download, then search the party names, the docket number, and one exhibit label before you remove the upload from the site. Open each hit and compare it with the page image. A pleading title that fails, or a Bates number that lands on the wrong sheet, means you keep the scan and convert the file again from a copy.

Two points are particular to a searchable PDF job:

• The file on your computer is untouched. The browser sends a copy, so the PDF you started with is still unsearchable. The searchable one is a new file.
• You choose the output. A tool that offers several exports hands back a searchable PDF only if you pick that format. Word or plain text holds the same recognized words without the page pictures.

From where you sit, Schema Law has the same outline. You sign in with Google, upload a PDF, PNG or JPEG from the browser, see the parsed result page by page, and export. The how it works section of the homepage puts it in three steps and says of the second that you can check the result against the original before you export.

Two things are true at the end that were not true at the start. With a tool that reads on its servers, a copy of your document is now held by the provider. And a new file is sitting in a folder on your computer that is not the matter file. Both need dealing with before the job is done. Whether a browser tool is the right kind of tool for your office at all is the subject of the post on online OCR versus desktop OCR for law firms.

Which PDFs should not be made searchable on a website?

Keep a PDF off a website when an order, an agreement or a client instruction limits where it may go, when you cannot find out what the site does with uploads, when the file carries a restriction its author set, or when it does not need OCR in the first place.

Run through this list before you upload.

The first three rows are judgments about confidentiality, and this post does not make them for you. The post on whether it is safe to upload client documents to an OCR tool lists the questions to put to a vendor before any client file goes up.

For reference, the lines Schema Law lists for every plan include four commitments: "Your documents are never used to train an AI model", "Documents stay inside your own account", "Encrypted in transit and at rest" and "Delete any document, at any time". Those words are from the list on the pricing page, and the security section of the homepage explains each one. Whether those commitments satisfy the terms that bind a particular document is for you to decide.

A test worth doing with any new site: upload a page with no client content first. A blank form or a page from a public court opinion shows you the whole flow, including what the tool keeps afterwards, at no risk.

Upload limits that stop an online searchable PDF job

An online job stops when the file is too big, has too many pages, is the wrong type, or would take you past your plan's allowance. Find all four numbers before you start. Schema Law accepts PDF, PNG and JPEG files of up to 20 MB, with a page cap per document that depends on the plan.

The Schema Law figures

The 20 MB upload size and the three accepted file types are the same on every plan. The pricing page has the prices and the rest of each plan.

Checking a file against the limits

• Size. Look at the file's size in your file manager. Color scans at high resolution are large. A long color scan can exceed a megabyte limit where the same pages in black and white would not.
• Pages. Open the file and read the page counter. Compare it with the per-document cap, not the monthly allowance.
• Type. A TIFF, a Word file or a ZIP of images is not a PDF, PNG or JPEG. Convert or unpack it first.
• Allowance. If the month's pages are nearly used, a long file may not fit.

When the file is over

Split an oversized PDF at document boundaries, such as between exhibits, and convert the parts. Keep a note of the parts and their page ranges so the page count can be reconciled afterwards. If the file is over on megabytes only, splitting solves that too, and it avoids recompressing the page pictures. If your routine files exceed a plan's cap, the plan is the wrong size. If the aim is to stay within a free allowance, the post on making a PDF searchable for free works out when that stops paying.

Downloading, naming and filing the searchable copy a website returns

The searchable copy a website returns lands in your browser's downloads folder under a name the site chose. Move it into the matter file the same day, rename it so it cannot be mistaken for the original, and confirm it opens and searches before you close the browser tab.

A sequence that works:

1. Download once. Repeated clicks produce file (1).pdf and file (2).pdf, and later nobody knows which is which.
2. Open the download and run one search for a word visible on the last page. If it fails, export again while the document is still on the site.
3. Compare the page count with the file you uploaded.
4. Rename it. Keep the original's name and add a suffix, for example 2024-03-11 Lease Tarrant-Bexley_ocr.pdf.
5. Move it from downloads into the matter folder, next to the original. Move, do not copy, so nothing is left behind.
6. Note it. One line in the file: which tool, which date, which pages you checked.

The downloads folder is outside your document system, and it may sync to a personal cloud account. Empty it of client documents at the end of each session. Finding a download that went astray and deciding which copy is current are covered, for a Word export, in the post on online OCR from PDF to Word under where the downloaded file went.

For how far to test the copy before others rely on it, see the post on checking whether a PDF is searchable.

Deleting the uploaded PDF once the searchable copy is saved

Delete the upload from the website once the searchable copy is saved and checked, unless you have a reason to keep it there. A document left on a web tool is a copy of a client's file held outside the matter file, and an extra copy to account for when the matter closes.

On Schema Law, deletion is yours to do at any time. The homepage states: "Delete any document whenever you choose. The original file and everything parsed from it are removed." No plan deletes documents for you after a set time, so do not count on an upload disappearing on its own. The post on whether OCR software stores your documents covers what a tool keeps and for how long.

For any other site, look for three things in its own terms or help pages: whether you can delete a file yourself, whether files are removed automatically and after how long, and whether deletion covers the result as well as the upload. If the site does not say, treat that as an answer.

A closing checklist for each job:

• The searchable copy is in the matter folder and opens.
• One search on a late page succeeds.
• The original as received is still in its place, unchanged.
• The downloads folder holds no copy.
• The document is deleted from the website, or its retention is noted in the file.

There is one reason to wait before deleting. A document you have uploaded to Schema Law can be exported again in another format, so if a Word or plain text copy may be wanted next week, keeping the upload saves reading the pages twice. Decide that on purpose, write down the date you will delete it, and then do it.