To get the text out of a screenshot or a photo of a document, add the picture file to MD Factory. It recognizes the text on your Windows PC and saves a Markdown file next to the picture, with the picture at the top and the recognized text below it. PNG, JPG, TIFF, WebP, BMP and GIF files work.
The picture stays in the same file because OCR makes mistakes, and you can check the text against the image without opening a second window. This guide covers standalone picture files. For pictures inside a PDF or Word file, see the links further down.
Convert a picture
- Open MD Factory and add the picture, or drag it onto the window. To take every supported file in a folder, add the folder.
- Conversion starts right away. Wait until the row shows Completed, Needs review or Failed.
- Open the Markdown file next to the picture.
Each picture gets its own Markdown file in the same folder, with the same name. Your picture file is not changed. If you convert the same picture again, MD Factory replaces its own earlier result only when you have not edited it. Otherwise it saves a numbered copy, such as receipt (2).md, so your edits stay.
The text is recognized on your PC with local OCR. The picture is not uploaded.
What the Markdown file contains
The file starts with the picture, followed by the text recognized in it. A copy of the picture is saved beside the Markdown file, and the file refers to it.
Move the Markdown file and the saved picture together. If you move only the .md file, the picture at the top no longer shows.
The Markdown holds the recognized text, not the look of the page. Colors, fonts and the position of text on the picture are not kept; the picture at the top is there for that.
Supported picture formats
| Format | What to know |
|---|---|
| PNG | The usual format of screenshots. |
| JPG, JPEG | The usual format of photos. |
| TIFF, TIF | A multi-page TIFF, such as a fax or a scan, gives one section per page. |
| WebP | Converted like the others. |
| BMP | Converted like the others. |
| GIF | Only the first frame is read. An animation is not played through. |
HEIC and HEIF files, the default photo format on iPhones, are not supported. Convert the photo to JPG first, then add the JPG.
Photos of documents
MD Factory applies the orientation that the camera saved in a photo, so a photo your phone marked as rotated is read upright. You do not have to turn it first.
Recognition quality still depends on the picture. MD Factory uses the Tesseract OCR engine, and Tesseract’s guide to improving the quality of the output names what helps:
- Resolution of at least 300 dpi. The guide says Tesseract works best at that resolution or higher.
- Text lines that run horizontally. The guide says line segmentation gets much worse when a page is too skewed.
- No dark borders around the page. The guide says scanned borders can be picked up as extra characters.
- Dark text on a light background. For Tesseract 4 the guide says to use dark text on a light background.
The last point matters for screenshots of dark-theme apps and websites. If the result is poor, take the screenshot again with a light theme.
Very large pictures are scaled down for recognition. If a small line of text in a big picture comes out wrong, crop the picture to the part you need and convert the crop.
Check the recognized text
The row ends with one of three statuses:
- Completed: text was recognized and the app found nothing to flag.
- Needs review: the file is saved, but the recognition confidence was low, or no text was found in the picture.
- Failed: the picture could not be converted.
A picture with no text, such as a photo of a landscape, still converts. You get the Markdown file with the picture and no text, marked Needs review.
Completed does not mean every character is right. Compare the numbers, dates, names and codes you will use with the picture at the top of the file. A wrong digit looks as plausible as a correct one.
The OCR language can be left on automatic or chosen. If a result is Needs review and you know the language of the text, choose it and convert again.
Which method to use
| Method | Best for | Limits |
|---|---|---|
| Snipping Tool, Text actions | One screenshot. Select the text or choose Copy all text, then paste it. | One image at a time. The text goes to the clipboard, not to a file. |
| MD Factory | Picture files and whole folders of them, with each result saved as a Markdown file beside the picture. | The result needs a check against the picture. |
Microsoft says Snipping Tool runs text recognition locally on your device. For a quick copy of a few lines it is the faster route. Use MD Factory when you have many pictures, want a file you can keep, or want the picture and the text together.
To convert a folder of screenshots, add the folder. Every picture gets its own file. To merge the results into one, see how to combine documents into one Markdown file.
Pictures inside other files
This guide is about picture files. Other cases have their own guides:
- A scan saved as a PDF: how to convert a scanned PDF to Markdown.
- A PDF where you are not sure whether the text can be selected: how to extract text from a PDF.
- Saving the pictures that sit inside a PDF, Word or PowerPoint file: extract images from documents.
Questions
Can I convert iPhone photos?
Not in HEIC or HEIF format. Convert the photo to JPG first.
What happens to a multi-page TIFF?
It becomes one Markdown file with a section for each page.
Does MD Factory upload my pictures?
No. Recognition runs on your PC. If the folder syncs to a cloud service, the Markdown file and the saved picture sync with it. The MD Factory Privacy Policy explains how the app handles your documents.
Does it change my original picture?
No. The picture stays as it is, and the result is a separate file.