Schema LawSign in with Google

Google Lens Copy Text: What It Does With a Document

By Schema LawPublished Updated

To copy text from a document with Google Lens, open Lens in the Google app, photograph the page or pick a photo, tap the text, tap Select text, then Copy. What you end up with is the selected words on your phone's clipboard. Google's help pages describe no saved file, no page order and no export.

That is a good fit for one clause or one address, and a poor fit for a 30-page record. This post follows Google's own help pages for every statement about Lens, and marks the places where those pages say nothing, so you know what to test on your own phone.

How do you copy text from a document with Google Lens?

Open the Google app, tap the Lens icon in the search bar, choose Search under "Search with your camera", then take a photo of the page or pick one you already have. Tap the text, tap Select text, and choose Copy from the menu. The words go to the clipboard, ready to paste.

Those steps come from Google's help article on region selection in Lens, which gives the same sequence for Android and for iPhone and iPad. Laid out as a routine for a paper document:

StepWhat you doWhat Google's article says at that point
1Open the Google app and tap Lens in the search barThe entry point on both platforms
2Under "Search with your camera", tap SearchLens has other modes there, including Translate and Homework
3Tap the shutter, or choose a photo"Take or upload a photo for your search"
4Tap the text, or drag the corners of the box over itThis is how you "find the 'Select text' button"
5Tap Select textA menu appears around the text
6Tap Copy, then paste where you need itCopy is the first action listed

Getting the selection right

Google's article explains the box in a few words: you can drag its corners, and you can move the whole box by tapping and holding its middle. For a legal document, take a moment over the edges of the selection. A clause that begins mid-line or ends before its proviso is worse than no quotation at all. After you paste, read the first and last words against the paper.

Before you take the photo

Google's only quality advice on that page is to "ensure that you use a clear and well-lit image." Flat paper, even light and a camera held square to the page all help any recognition tool. The guide to taking a photo that OCR can read covers the technique in detail.

A worked example

An associate at a fictional firm, Rennick Vale, is at a client's warehouse and is handed a signed delivery receipt. She needs the consignee's address in an email to the office right away. She opens Lens, photographs the receipt, taps the address block, taps Select text, taps Copy, and pastes into the email. She then reads the street number and postcode against the paper before sending, and photographs the whole receipt separately for the file.

What the Google Lens Select text menu offers besides Copy

Besides Copy, Google's article lists Listen, Translate and Search, followed by actions that depend on what you selected: email an address, add a number or email to contacts, call or text a phone number, open an address in Maps, open a link, or track a package number.

Here is the full list from Google's region selection article, grouped by when each action appears.

ActionWhen Google says it is offeredA use in legal work
CopyListed for selected textPaste a passage into an email or draft
ListenListed for selected textHear a paragraph read aloud while you follow the paper
TranslateListed for selected textGet the gist of a foreign-language notice
SearchListed for selected textLook up a company name or a statute citation
Send an email"if the email address is selected"Reply to the address printed on a letterhead
Add to contacts"if the phone number or email is selected"Save an adjuster's details from a claim letter
Call, Text"if the phone number is selected"Phone the clerk's number from a notice
Maps"if the address is selected"Find a hearing location
Open a website"if the link is selected"Open a portal address printed on a form
Track a shipment"if a package number is selected"Check a courier reference on proof of service

Google's article also has a tip about Listen: once you choose it, a Stop control is shown until the reading ends or you stop it.

Cautions that belong with three of these

Translate. Google's Photos help says Lens "can translate text into all Google Translate languages." A machine translation is a way to learn what a document is about. Whether a translation is acceptable for filing is a question for your court's or agency's rules; check the rule that applies to you.

Search. Choosing Search sends the selected words to Google as a query. Think before you do that with a client's name next to the subject of the matter.

Listen. Text read aloud in a courthouse corridor is heard by whoever is standing there.

Google Lens on Android, on iPhone and inside Google Photos

Google Lens lives in the Google app and inside Google Photos. Google's Photos help says that on Android you can use Lens from Google Photos, from Google Assistant on most phones and from the Google app on some phones. On iPhone and iPad it works with Google Photos or the Google app, "but not with your camera or Google Assistant."

The iPhone statement is on the iPhone and iPad version of the Photos help article; the same address without the platform setting can open on the Android version.

WhereAndroid, per GoogleiPhone and iPad, per Google
Google appThe region selection article gives the Google app route for Android devices. The Photos help article says the Google app offers Lens "on some Android phones, like Pixel"The region selection article gives the same Google app route
Google PhotosOpen a photo and tap LensOpen a photo and tap Lens
Google Assistant"on most Android phones"Not available
The phone's own camera appNot addressed in the articles opened for this postGoogle says Lens does not work with your camera

The two Google articles describe the Android Google app slightly differently, as the first row shows. If Lens is missing from the search bar on your phone, the Photos route is the one both articles agree on.

The Google Photos route

Google's article on how to get info about your photos and surroundings gives the Photos route: open the Google Photos app, select a photo, and tap Lens. For text, the article has a one-line tip: "To copy text, tap and select the words you'd like to copy."

This route suits a page you photographed earlier, or a picture a client sent you that you saved to the phone. One consequence follows from using Photos at all. If Google Photos backs up your camera roll to your Google Account, the photograph of the client's document is now stored there. That is a storage decision made by your backup setting, separate from anything Lens does. For the tools built into an iPhone or a Samsung phone, see the post on copying text in a photo.

On an iPhone

Since Google says Lens does not work with the iPhone camera, you need the Google app or Google Photos installed. For what Apple documents on its own devices, see the post on Apple's built-in document scanner in Notes.

What does Google say happens to an image you search with Lens?

Google says it processes the image with its visual search technologies. Whether the image is then saved to your Google Account depends on the Visual Search History setting, which Google says "is off by default". When that setting is on, images you search with "may be saved to your Web & App Activity" and may be used to train Google's models.

The source is Google's help article on how to manage your Visual Search History. The statements a lawyer should read in Google's own words:

When you use an image to search, Google uses its visual search technologies to process the image, like when you search with Google Lens.

If your Visual Search History setting is on, Google may use the images you use to search on eligible Google services to train the models that develop and improve Google's visual search technologies.

On what turning the setting off does and does not do, the article lists these points among others:

  • "This doesn't turn off Web & App Activity, which can still save your search activity on Google services."
  • "Previously saved images can still be used to improve visual search technologies."
  • "You can delete previously saved images in your Web & App Activity at activity.google.com."

The article adds that if a separate option called Visual Search History Donation is on, saved images "are eligible for review by trained human reviewers". It also notes that the Visual Search History setting "does not affect images that may be saved by other settings, like Gemini Apps Activity."

What the article does not tell you

The article is about what is saved to your account. It does not describe where an image is processed, how long Google holds an image that is not saved to your activity, or who at Google can reach it. This post makes no claim on those points. If the answer matters for a document, the article does not give it to you.

The page also opens with a notice that Google is "currently updating settings for Google Search services" and that the settings described may not match what you see. Read the live page on the day you check, on the account you use.

A settings check on your own account

  1. Sign in to your Google Account and open Data & privacy.
  2. Under History settings, open Web & App Activity.
  3. Look at the Visual Search History box. Google's article says checking or unchecking it turns the setting on or off.
  4. Open the Lens page on My Activity, which Google's Photos help names as the place to "find and delete your Google Lens activity".
  5. Do the same on every account signed in on the phone. A personal account and a firm account have separate settings.

What this means for a client document

A photograph of a client's document sent to a consumer service is a disclosure question before it is a technology question. Whether it is permitted is for your firm's policy, the client's instructions and your jurisdiction's professional conduct rules, and this post does not answer it. The post on whether it is safe to upload client documents to an OCR tool sets out the questions to put to any service that handles them.

Why Google Lens gives you a passage and not a document

Lens copies the words you selected to the clipboard, and that is all the help articles describe. They describe no saved file, no way to capture several pages as one item, and no export. So the copied passage carries no page number, no page image and no link to the paper it came from.

Three gaps follow from that for document work.

No file

A paste is text in whatever you pasted it into, and the photograph in your camera roll is still an image with no text in it. The Snipping Tool post explains clipboard text versus a text layer written into the file, and the guide to making a scanned PDF searchable explains how that layer is added.

No page order

Each capture is one photograph. Google's articles describe selecting text in one image at a time. For a ten-page agreement, that means ten photographs, ten selections and ten pastes, and you are the one keeping them in order. Nothing in the clipboard records that a paragraph came from page 7. Write the page number into your notes as you paste.

No tables

The clipboard holds text. Google's articles do not describe copying rows and columns as a table. If you select a schedule of payments, test what arrives when you paste into a spreadsheet: whether each figure lands in its own cell or the whole selection lands as lines of text. Decide from that result, on your own document, and expect to rebuild a table by hand if the paste comes through as a run of numbers.

Paste the selection into a note, then check it for the contract date, the invoice amount, and the notice address. If the photo shows an affidavit or an exhibit stamp, find that short string and read it on the picture before you forward the paste. Close the photo only after those details agree.

When a lawyer should stop using Google Lens and convert the whole scan

Stop when the document itself has to be kept, searched, cited, shared or filed. Lens serves a moment: you need a few words now. A matter needs a file: every page, in order, with text that can be searched and checked against the image months later.

The dividing line, by task

If the task isUse
Copy a short passage once and move onLens
Read or translate something on the spotLens
More than a page or two of textConvert the scan
Anything you will cite by pageConvert the scan
A document that goes into the matter fileScan it properly, then convert
A table of figuresConvert, with a table-capable export
A document that leaves the office as a searchable PDFConvert the scan
A privileged or sensitive documentFollow firm policy before any consumer app

Signs you have crossed the line

  • You are photographing the fourth page of the same document.
  • You have pasted from the same document twice today.
  • Someone asks you which page a quotation came from and you have to find the paper.
  • A colleague asks you to send them "the text".
  • You are retyping numbers because the paste broke a table apart.

What converting the whole scan involves

Get every page into one file first. A sheet-fed scanner does this in one pass. On a phone, a scanning app that saves a multi-page PDF is the tool, not a row of separate photos; the post on Genius Scan and its text recognition covers what that app's maker documents about recognition and PDF export.

Then run recognition on the file as a whole. In Schema Law you upload a PDF, PNG or JPEG in the browser, the pages are read into text and tables, and you export the result; the how it works section of the homepage shows the steps. Searchable PDF keeps the page images with the text behind them, Word gives you editable language, and the spreadsheet formats, Excel and CSV, hold tables. Every plan includes all of them.

Whichever tool does the reading, check it. Recognition is strongest on clean print and weaker on handwriting, and nothing marks a passage as uncertain, so read names, dates and amounts against the page.

All posts in OCR in everyday software