How to Extract Text From an Image on Android (4 Routes)
Android has text recognition in three different places, and which one you get depends on your phone's manufacturer and version. On Pixel it is built into the system; on Samsung it lives in the keyboard and the gallery; on other phones you may need Google Lens. Here is each route.
Route one: Google Lens
This is the one that works on essentially every Android phone, regardless of brand.
Open the Google app or Google Lens app, point the camera at text, and tap to capture. A text-selection overlay appears on the photo — you can select just the words you need, or tap "Select all" and copy.
The same thing works on screenshots: open the screenshot in Google Photos, tap Lens, then the text tab. Screenshots work better than camera captures here for the obvious reason that a screenshot has perfect edges.
Route two: the recent-apps screen on Pixel
On Pixel devices this is the fastest path and few people know about it.
Open your recent apps, then long-press on any text visible in an app's preview thumbnail. Android offers to select it. You can copy from an app that otherwise gives you no way to copy — useful for apps that display text but disable selection.
This does not require taking a screenshot first.
Route three: Samsung's keyboard
On Samsung phones, the Samsung Keyboard has a text-scanning button in its toolbar. Tap it, and you can scan text from the camera or from an image in the gallery, and it inserts the result straight into whatever you were typing in.
That last part is the advantage: the text lands where you need it rather than in the clipboard.
Route four: anything that works everywhere
If your phone is none of the above, or if the built-in routes are being awkward, a browser tab works identically on every Android device.
Open the screenshot to text tool in Chrome, drop or paste the image, and the text appears. It runs locally, so nothing is uploaded and it works on mobile data without a round trip. This route also has one advantage over the built-in ones: it can produce a .docx file rather than just clipboard text, if you use the image to Word converter.
What all of these do badly
Worth knowing before you conclude the tool is broken.
Columns. Every line-based reader — Lens, Bixby, or anything else — scans across the full page width. A two-column layout comes out interleaved. Crop one column at a time.
Cursive handwriting. Block letters read fine. Joined-up writing produces nonsense, on every platform, because the underlying problem (no clear letter boundaries) is the same everywhere.
Very small print. Below about 12 pixels per character there is nothing to work with. Get closer, or crop.
Text over a photo. Letters and background at similar brightness cannot be separated. No amount of tapping the screen fixes this.
Low-contrast dark mode. Pale text on near-black is workable but weaker. Switching the app to light mode and re-capturing often works better than any processing.
The quality differences that matter
It is tempting to assume one of these is "the good one". In practice they are all doing the same kind of recognition on the same pixels, and they differ in tuning rather than in capability.
What genuinely changes the result:
Whether it was a screenshot or a photo. A screenshot has clean edges and no noise. This is a bigger factor than which app you used.
Whether the image was recompressed. Anything that travelled through WhatsApp or WeChat has been shrunk. Use the original.
How you cropped it. Text filling the frame reads far better than text occupying a fifth of a wide shot.
Light and angle, if it was a photo. Shadow and tilt do more damage than any tool can repair.
If you need the text in a document
None of the built-in Android routes produce a file. They produce a clipboard entry. Pasting that into Google Docs loses paragraph structure — a page of text arrives as one block.
If the end goal is an editable document, extract the text and paste it into Docs line by line, or use a converter that builds the document for you. That is worth knowing before you spend twenty minutes reformatting after a paste.
On privacy
Lens processes on-device for the recognition step, but it is a Google product and the results interact with your Google account in ways that are not always obvious. The Samsung keyboard scanner similarly connects to Samsung services.
If the text is something you would not want associated with your account — a medical document, a bank statement, a work contract — the safest route on Android is a browser tab running local recognition, where the image never leaves the phone.