5 Ways to Extract Text from an Image (That Actually Work in 2026)

Last week I stood in a parking lot squinting at a rental car agreement, trying to type the confirmation number into my phone. Twenty-six characters of random alphanumeric garbage. I got four characters in before I thought: why am I doing this? I opened my camera, grabbed a photo, and had the text on my clipboard in about three seconds. That moment—swapping ten minutes of careful pecking for a quick snap—is the whole reason these tools exist.

You don’t need one magic tool. You need to know the options so you can pull text out of an image without wasting time or sending sensitive stuff to the cloud when you don’t have to.

Here are five ways that actually work—and what nobody talks about: where your image goes when you hit “extract.”

Why you’d want to pull text out of an image

Three situations I run into all the time:

1. You have a photo of a document. A receipt, a form, a page from a book. You need the text in a doc or spreadsheet so you can search it, edit it, or send it in an email. Retyping a 300-word receipt is not just slow—you will make mistakes. I once mis-typed a deductible amount on an insurance form because I read a “3” as an “8.” That was a fun phone call.

2. You’re looking at a screenshot. An error message, a tweet, a recipe from a story. You want to copy the text, not retype it. Sometimes the app won’t let you select it, so the only way is to grab it from the image. I run into this constantly with screenshots of code snippets people send me on WhatsApp—you can’t copy code from a JPEG.

3. You’re dealing with handwriting. Meeting notes, a signed form, old letters. You want them searchable or in a note-taking app. That’s harder than printed text, but doable with the right method. My handwriting is somewhere between “rushed doctor” and “ancient scroll,” so I have some experience pushing OCR to its limits.

If any of that sounds familiar, the rest of this will map directly onto your use case.

Method 1: Your iPhone already does this (Live Text)

Live Text is built into iOS (starting with iOS 15, so any iPhone from the last few years). You see text in a photo, in the camera, or on a webpage—you tap and hold, then copy. No app to install, no account, no upload. It’s free and works offline.

I use it probably five times a day without thinking about it. WiFi passwords on stickers, serial numbers on electronics, street addresses on signs when I need to paste them into Maps. The whole interaction takes about two seconds: tap, hold, copy. Done.

Pros: Fast, private (on-device), good for short bits of text. Zero setup. Cons: Not great for long paragraphs, handwriting, or exporting. Sometimes it doesn’t offer the “copy” option at all on messy layouts. I once tried to grab a full menu from a restaurant photo—it kept selecting one line at a time and missing the prices on the right column.

When I use it: Single lines, WiFi passwords, quick quotes, anything where “copy and paste” is the only goal. If I need more than a couple of lines, I move to something else.

Method 2: A dedicated OCR app (e.g. Textora)

An OCR app is built to extract text from images. You take or pick a photo, it runs OCR, and you get editable text you can copy, export, or save. Many of them run on the device, so your image doesn’t have to leave your phone. Textora is one of them: it handles printed and handwritten text, supports many languages, and lets you export or organize the result.

The difference between Live Text and a proper OCR app shows up the moment you have more than a few lines. A full receipt with 30 line items, a contract page with fine print, or a handwritten note from a meeting—that’s where dedicated apps earn their place. They process the whole image at once and give you a clean block of text you can do something with.

I tested this last month with a two-page lease agreement someone photographed and sent me. Live Text got about 70% of it with some line-break chaos. The OCR app got 95%+ with the layout mostly intact. That’s the gap.

Pros: Better accuracy on long text and handwriting, batch processing, export options, often works offline. Cons: You have to open an app and go through a few steps. For one sentence, that’s more than you need.

When I use it: Full pages, receipts, handwritten notes, or when I need a clean block of text to paste into Notion or a report. I wrote a more detailed image-to-text workflow on iPhone if you want the exact steps.

Method 3: Google Lens

Point your camera at text or upload a photo; Lens can copy the text or overlay a translation. Handy when the text is in another language or you’re in a hurry and already have Google open.

I travel a fair amount, and Google Lens is the tool I pull out at restaurants in countries where I can’t read the menu. Point, translate, order. That workflow alone makes it worth having. The real-time overlay where it replaces the foreign text with your language right on the screen is genuinely impressive—I used it on a Japanese train schedule last year and it saved me at least an hour of confusion.

For English-only printed text, though, it’s not my go-to. The output is a bit unstructured—you get the text but not in a clean format—and it really wants to be online. If I have no cell signal, I have no Lens.

Pros: Good for quick lookups and translation, no extra app on many Android phones, real-time camera overlay. Cons: Tied to the internet for translation and full features, not ideal for long documents or structured export. Results can be messy for dense text. Doesn’t give you a “save” or “export to file” flow.

When I use it: Menus, signs, and labels in another language. Quick “what does this say?” moments. For English-only or long text, I usually use Live Text or an OCR app.

Method 4: Online OCR tools

You upload an image to a website, it runs OCR in the browser or on their server, and you download or copy the text. Fast and no app install. You’ll find dozens of these with a quick search: OnlineOCR, i2OCR, Free-OCR, and others.

The privacy thing nobody talks about: Your image is uploaded to someone else’s server. That might be fine for a random flyer or a concert poster. It’s not fine for a contract, an ID, a medical slip, or a receipt with your card’s last four digits. I made this mistake once—uploaded a photo of an insurance card to some OCR site because I was lazy, then realized I’d just sent my policy number, group number, and name to a server I knew nothing about. Not great.

Read the privacy policy. If you can’t tell where the image is stored or how long it’s kept, assume it’s stored forever. For sensitive stuff, use an on-device method (Live Text or an app that processes locally). This is the single biggest thing I wish someone had told me earlier about online OCR tools.

Pros: No install, works on any device with a browser, often free for a few pages. Cons: Your image leaves your device. Not ideal for sensitive or personal documents. Some sites add watermarks, limit the number of pages, or require sign-up for full results.

When I use it: Only for non-sensitive, one-off stuff when I’m on a computer without an OCR app. A poster, a flyer, a public sign. Never for anything with personal data.

Method 5: Desktop apps (e.g. ABBYY FineReader)

Full-blown desktop OCR software can do batch jobs, complex layouts, and multiple languages. They’re powerful and overkill for most people. ABBYY FineReader, for example, can handle a 200-page scanned PDF and turn it into a fully searchable, editable Word document with the original formatting mostly intact. That’s impressive—but it’s also $200+ and requires a desktop.

These apps shine in very specific scenarios: you’ve inherited a box of scanned documents from an office that closed, or you need to digitize an entire filing cabinet. If you’re a law firm, an accounting team, or an academic researcher dealing with thousands of pages, desktop OCR is still the gold standard.

Pros: Maximum control, batch processing, advanced options, handles complex multi-column layouts. Cons: Cost, setup, and you’re tied to a computer. For “I have a photo of a receipt on my phone,” this is rarely worth it. Learning curve is real.

When I use it: Almost never for everyday image-to-text. Only when I have a big, formal digitization project—like the time I scanned 80 pages of old family recipes from a handwritten notebook my grandmother kept. That was a desktop-OCR-plus-manual-correction job.

Side-by-side: which method for which situation

Situation	Method I’d use
One line from a photo	Live Text
Full page, need export	OCR app (e.g. Textora)
Handwriting	OCR app
Text in another language	Google Lens
Sensitive document	Live Text or OCR app (on-device only)
Quick one-off on PC	Online OCR (non-sensitive only)
Large batch (100+ pages)	Desktop OCR software

Tips for getting cleaner text from messy images

Even the best tool can’t save a terrible photo. Here’s what makes the biggest difference:

• Crop so the text fills the frame. Less background = fewer mistakes. If you can cut out the watermark, the logo, and the decorative border, do it. I’ve seen accuracy jump from ~80% to ~95% just by cropping tightly.
• Lighting: Even, no glare. Shadows and reflections confuse OCR badly. The worst offender? Overhead lights reflecting off glossy paper. Tilt the page slightly or move to diffused light.
• Angle: Shoot straight-on. Angled shots stretch or distort characters. Even 15 degrees off can turn an “l” into a ”/” in the output. Hold the phone parallel to the page.
• Resolution: Blurry or tiny text = more errors. Get close or use a higher-res image. If the text is smaller than about 10pt equivalent in the photo, most OCR will struggle.
• Contrast: Dark text on light background works best. Pencil on yellow paper is notoriously hard. Blue pen on white is usually fine. Red text on pink? Good luck.
• Language: Tell the app which language(s) to expect. Wrong language = weird substitutions. I once got an entire paragraph of accented characters because the app was set to French while reading English text.

If your results are still bad, this troubleshooting guide goes through the most common fixes step by step.

The real decision is about privacy

Most of the time, the accuracy difference between these methods is small for clean printed text. The bigger question is: where does your image go? Live Text and on-device OCR apps keep the image on your phone. Online tools and some cloud-based apps send it to a server. Desktop apps stay on your computer.

For a random flyer, it doesn’t matter. For your ID, your medical records, or a photo of your kid’s school form—it matters a lot. Pick the method that fits both your situation and your comfort level with where that image ends up.

Bottom line: you have several ways to extract text from an image. Pick the one that fits your situation and your privacy comfort—and when in doubt, keep the image on your device.