Lee este artículo en español.

OCR in Make: Read Text from Images (Built-in or API)

Short answer: Make already includes OCR. Use the built-in Make AI Content Extractor app, module Extract text from an image (printed or handwritten text, 2 Make credits per operation). Google Cloud Vision is another option if you have a Google Cloud account. If you want the image read by Tesseract hosted by Cofferdock, and not sent to a third-party AI, you can call Cofferdock's Image OCR from the HTTP module. The same call works in n8n, Zapier or your own code.

If you need to read the text in a photo, a scan or a screenshot inside a Make scenario, you have three real options. This guide starts with the built-in way, then shows Google Cloud Vision and the Cofferdock API, with step-by-step instructions for the HTTP module.

The built-in way: Make AI Content Extractor

1. In your scenario, add the Make AI Content Extractor app.
2. Choose the module Extract text from an image.
3. In the image field, map the image that comes from a previous module (for example, a file from an earlier step in your scenario).

Good to know:

• It reads printed or handwritten text.
• Each operation uses 2 credits of your Make plan.
• Images must be under 20 MB.

Other option: Google Cloud Vision

Add the Google Cloud Vision app to your scenario and choose the module Run Text Detection (OCR) within an Image. You need a Google Cloud account to use it.

When Cofferdock's Image OCR is worth it

• The image is not sent to a third-party AI provider. Cofferdock runs Tesseract on its own servers.
• It reads English and Spanish.
• It works very well on clean printed text, such as invoices, receipts and printed pages.
• The same API works in Make, n8n, Zapier and your own code.
• Each image costs 1 Cofferdock credit.

To be plain about it: for handwriting, Make's extractor or a specialised service is the better pick. Cofferdock's OCR works best on clean printed text.

Step by step with the HTTP module

1. Add the HTTP app to your scenario and choose Make a request.
2. Fill in the fields:

- URL: https://api.cofferdock.com/image-ocr - Method: POST - Headers: add x-api-key and paste your Cofferdock key here. Keep the key in the header, never in the URL. - Body type: Raw - Content type: JSON (application/json) - Request content:

{
  "file": "{{toBase64(1.data)}}",
  "language": "eng"
}

Use spa instead of eng for Spanish text. Here 1 is the number of the module that gives you the image.

1. Turn on Parse response so Make reads the JSON answer.

After that, the text is in the field text (for example {{2.text}}, where 2 is the number of the HTTP module), and the confidence of the reading is in confidence, a number from 0 to 100.

Scanned PDFs

Scanned PDFs have no real text inside, so they need a different step:

• Make a scanned PDF searchable (pdf-ocr): send the PDF to https://api.cofferdock.com/pdf-ocr. It returns a new PDF with an invisible text layer, so you can search and copy the text. It costs 1 credit per page, up to 30 pages per call.
• Then use Extract text from a PDF (https://cofferdock.com/pdf-extract-text) to get the text as plain text.

A quick example

Say you have invoices as photos in a Google Drive folder, and you want the text in a Google Sheets table:

1. A Google Drive module gives you each image in the folder.
2. The HTTP module sends each image to Image OCR, as described above.
3. A Google Sheets module adds a row with the file name, the text and the confidence.
4. Add a filter on confidence. Rows below a level you choose (for example, 80) go to a sheet tab you review by hand.

This gives you the text, not structured fields. Cofferdock does not pick out the total or the date for you; you read those from the text.

Prefer to see it in code?

curl -X POST "https://api.cofferdock.com/image-ocr?language=eng" \
  -H "x-api-key: YOUR_KEY" \
  -H "Content-Type: image/jpeg" \
  --data-binary @invoice.jpg

The answer looks like this:

{ "success": true, "text": "INVOICE #2026-001\nTotal: $145.00", "confidence": 94, "meta": { "language": "eng", "used": 12, "remaining": 488 } }

How much does it cost?

• Cofferdock: 1 credit per image (PDF OCR uses 1 credit per page). The free plan gives you 50 credits a month, with no card. Failed calls are not charged. Paid plans start at €9 a month for 500 credits.
• Make AI Content Extractor: 2 credits per operation, from your Make plan.
• Google Cloud Vision: priced by Google. Check their pricing page before you choose it.

FAQ

Does Make have built-in OCR? Yes. The Make AI Content Extractor app has the module Extract text from an image.

Can it read handwriting? Make's extractor handles printed and handwritten text. Cofferdock's Image OCR works best on clean printed text, so for handwriting use Make's extractor or a specialised service.

Which languages does Cofferdock support? English (eng, the default) and Spanish (spa). No other languages for now.

Is my image sent to an AI company with Cofferdock? No. Cofferdock runs Tesseract on its own servers, and the image is not sent to a third-party AI provider.

Related guides

• OCR in n8n
• Extract text from a PDF
• Web scraping in Make

Try it yourself

Try the Image OCR tool at cofferdock.com/image-ocr. You can start free at cofferdock.com/register.