How to convert handwriting to text in Google Docs

You have a stack of handwritten notes, a letter, or a multi-page document, and you want the text in Google Docs so you can search it, edit it and share it. Google gives you three free routes (Drive, Keep and Lens), and all three run the same OCR engine. The question isn’t whether they work; it’s how well they work on your handwriting.

Here’s the honest answer: Google’s OCR is good at printed text and acceptable at neat block-letter handwriting. On normal cursive it drops to around 40% to 60% word accuracy, which means a typical page comes back with dozens of broken words, garbled proper names and missing lines. This guide walks through every Google route with real accuracy numbers, then covers what to do when the free tools aren’t accurate enough, including the Handwriting OCR add-on that brings a handwriting-trained engine into Drive and Gmail.

Quick takeaways

• Google Docs has no built-in handwriting OCR. The “OCR” happens in Drive when you open an image or PDF with Google Docs, and it uses Google’s general OCR engine.
• Accuracy is fine on clean printed handwriting (80% to 90%) but typically drops to 40% to 60% on cursive and below 30% on historical or messy scripts.
• Google Keep, Lens and most Workspace OCR add-ons all use the same underlying engine, so switching between them rarely improves accuracy: it just changes the workflow.
• File limits matter: Drive OCR caps at 2 MB images and 10-page PDFs. Beyond that, output is silently truncated.
• For cursive or any handwriting that matters, a dedicated handwriting OCR returns 95%+ on the same input and exports to Word, PDF or plain text in one click.
• If you’d rather not leave Drive at all, there is now a Handwriting OCR add-on on the Workspace Marketplace that runs the same engine inside Drive and Gmail.

Quick decider: which Google tool fits your situation

Google’s built-in OCR options

Google has OCR in three products you probably already use, plus the Workspace Marketplace. Each one has the same underlying engine but a different workflow.

1. Google Drive: “Open with Google Docs”

Google Drive can OCR an image or PDF and open the result as a Google Doc.

Step by step:

1. Upload the file. Drag and drop the JPG, PNG or PDF into Google Drive, or use New > File Upload. Make sure the image is right-side up. A quick rotate or contrast bump beforehand often improves the OCR result.
2. Right-click > Open with > Google Docs. Google runs OCR and creates a new Google Doc.
3. Use the result. The new doc contains the original image at the top and the extracted text below. The text is editable and exportable like any other Google Doc.

What it’s good at: Free, no installation, completely inside the Google ecosystem. Handles both images and PDFs (up to 10 pages). Preserves basic formatting in many cases.

What it isn’t good at: Cursive handwriting. The engine was built for printed text and treats connected letters as noise. Expect 40% to 60% accuracy on everyday cursive and below 30% on faded historical writing. File limits also bite: images larger than 2 MB and PDFs longer than 10 pages are silently truncated.

2. Google Keep: “Grab image text”

Google Keep has a built-in OCR command called Grab image text that’s handy for one-off notes from your phone.

Step by step:

1. Add the image to a Keep note. On web, use New note with image; on mobile, snap a photo or pick an existing image.
2. Open the three-dot menu and pick “Grab image text”. Keep extracts text and pastes it into the note below the image.
3. Send to Docs. From the same three-dot menu, Copy to Google Docs creates a new doc with the text already in it.

Strengths: Mobile-friendly, no separate upload step, free, and the “Copy to Google Docs” shortcut saves you the manual paste. Works on JPG, PNG, GIF and WebP up to 10 MB.

Limits: No PDF support. One image at a time. Plain-text output only, so any structure (columns, headings, indentation) is lost. Accuracy on cursive matches Drive: around 40% to 60%.

For a deeper walkthrough of the Keep flow, see how to convert handwriting to text in Google Keep.

3. Google Lens (and Google Photos)

Google Lens is the engine behind Google Photos’ text-recognition prompts and works on the fly from a phone camera or web photo.

On mobile: Open the photo in Google Photos, tap the Lens icon, and pick Copy text when it appears. Paste into Google Docs.

On desktop: Open the image at photos.google.com. If the Copy text from image prompt doesn’t surface automatically, right-click the image in Chrome and choose Search image with Google, then switch to the Text tab in the Lens sidebar.

Where Lens shines: Quick, ad-hoc captures. Pointing your phone at a recipe card and copying the ingredients to a doc. It supports many languages and handles mixed print plus cursive better than Drive in some cases.

Where it falls short: No batch mode. Each page is a separate copy-paste cycle. Connected cursive and faded historical writing still produce errors. Long documents are tedious.

4. Google Workspace Marketplace OCR add-ons

Search the Marketplace for “OCR” or “image to text” and you’ll see dozens of add-ons promising in-Docs OCR.

The truth about most of them: they call Google’s own OCR engine in the background and wrap it in a Docs sidebar. The convenience is real (you can stay inside the document); the accuracy improvement on handwriting is not. If Google Drive OCR can’t read your cursive, neither can the typical add-on.

So the question to ask before installing one isn’t whether it runs inside Docs. It’s whose OCR it calls when it does. An add-on that brings its own handwriting-trained engine changes the result; an add-on that re-skins Drive’s engine changes only the number of clicks.

What Google’s OCR actually returns on cursive: a real example

Here is the result of opening a clear, legible 1989 handwritten letter directly with Google Docs.

The first few lines of the output:

191 Strand Road Dublin q 5th April 1989 isii Dear Richard, 17 Busk Forgive the Melay in responding to your Frill manuscript.

Things to notice:

• Proper nouns broken or invented. “Dublin 9” became “Dublin q”. The letter writer’s name became “17 Busk”.
• Common words substituted. “delay” became “Melay”. “Full” became “Frill”.
• Line structure collapsed. Lines fragmented mid-sentence. Sentence boundaries lost.
• Whole phrases dropped. Some sections are simply missing.

A human reader handles this letter without trouble. Google’s OCR returns something that requires a complete rewrite, not a proofread.

When to use a dedicated handwriting OCR instead

Dedicated handwriting OCR services are trained specifically on handwriting (cursive, print, mixed, historical) rather than retrofitting an engine designed for typeset text. The accuracy difference on the same letter above is roughly 96% versus 45%.

A practical workflow that’s free to try:

1. Upload your handwritten image or PDF to Handwriting OCR. Free trial credits, no card required.
2. Pick Extract full text as the action.
3. Wait 15 to 30 seconds per page.
4. Download as Word, PDF or plain text, or copy the text and paste it into Google Docs.

You stay inside the Google ecosystem for everything except the OCR step itself, and you sidestep the typical 30 to 60 minutes of manual correction.

If you’d rather not leave Drive: the Workspace add-on

The same engine is available as a Handwriting OCR add-on on the Google Workspace Marketplace, which removes the upload and paste steps entirely.

Select a scanned PDF or image in Drive and the add-on panel picks it up. You grant access to that one file, connect your Handwriting OCR account with a browser sign-in on first use, choose an action and an output format, and the result is saved to your Drive next to the file you started from. In Gmail, open a message with a scanned attachment and the add-on lists that message’s attachments so you can process one without downloading it first.

The accuracy is identical either way. The add-on just saves the round trip, so it’s worth having if most of your documents already live in Drive.

What the add-on can return

The practical limits, plainly: PDF, JPG, PNG, TIFF, BMP, GIF, WEBP and AVIF, up to 20 MB per file. Processing costs 1 credit per page, drawn from the same balance as the web app. Tables and custom extractors need a Pro or Business plan; transcription and translation don’t. For archives beyond a handful of files, or for anything over 20 MB, the web app takes files up to 100 MB and handles batches in one go.

For the full walkthrough, including the Gmail flow and what the add-on can and cannot see, read the launch post.

On privacy: the add-on can’t browse, search or read your Drive. It sees the single file you select and the attachment on the message you have open, and nothing else. That’s a deliberate design choice, and it’s why the permissions prompt asks for so little. Details are in our privacy policy.

Where this matters most:

• Cursive handwriting, where Google’s OCR drops to 40% to 60%.
• Historical letters, where ink fade and script style compound the problem.
• Multi-page archives (more than 10 pages), where Drive truncates and Keep can’t help.
• Foreign-language handwriting, where you also need translation in the same pass.
• Names, dates and proper nouns, which Google’s OCR substitutes more often than it admits.

Tips that improve accuracy in any tool

These apply whether you use Google’s OCR or a dedicated tool. They’re worth a minute of preparation because they often turn a marginal scan into a clean one.

• Light the page evenly. Natural daylight or a single lamp at a low angle. Avoid direct overhead light on glossy paper (causes glare).
• Hold the camera parallel to the page. A square photo OCRs much better than an angled one. Phone document-scanner modes do this perspective correction automatically.
• Use dark ink on light paper. High contrast helps every OCR engine. Pencil or faded ink can drop accuracy 10 to 20 percentage points.
• Scan at 300 DPI, not 600. We render each page at up to 2000 pixels on the long edge, and an 11-inch page reaches that at roughly 180 DPI, so 300 is already comfortable headroom. Going higher just makes a slower, heavier file that gets scaled back down. See scanning handwriting to text for the full explanation.
• Split long PDFs. Google Drive caps at 10 pages. Even outside Google, splitting reduces the chance of a partial-failure run.
• Print is easier than cursive. If you’re taking notes for future OCR, write in block letters. It’s the single biggest accuracy win you can make at capture time.
• Proofread proper nouns specifically. OCR errors cluster around names, dates and unfamiliar words. Scan-edit those first.

Bottom line

For occasional, neatly-printed handwriting, Google Drive’s “Open with Google Docs” is the simplest free path and the answer is genuinely good. For cursive, multi-page archives, foreign-language documents, or anything where you can’t afford a 40% to 60% error rate, you need an engine built for handwriting. The difference now is that you don’t have to leave Google to use one.

Install the add-on and run it on a single page of your hardest sample, or try the web app if you’d rather start outside Drive. If the output matches what you can read by eye, the same will hold for the rest of the archive. If you have specific questions about a tricky document or a large batch, get in touch with a sample and we’ll tell you what to expect before you start.