Batch OCR for folders of PDFs

Batch OCR processes many PDFs without opening and exporting each one. It writes one UTF-8 .txt file per PDF. It needs the Google Cloud setup and a signed-in account.

In the desktop app

Choose Batch OCR PDFs on the welcome screen, or File → Batch OCR PDFs… (Ctrl+Alt+O).

1. Select PDFs… to pick files, or Use Folder… to take every PDF in a folder.
2. Choose the Output folder. It defaults to a txt folder beside the first PDF.
3. Keep OCR language as my for Burmese, or change it.
4. Click Start OCR.

A progress dialog shows each file. You can cancel at any time. If your Google sign-in has expired, LEXO offers to sign in and continue.

From the command line

A folder (every PDF directly inside it):

lexo ocr-batch ./pdfs --out-dir ./txt

Specific files:

lexo ocr-batch chapter-a.pdf chapter-b.pdf --out-dir ./txt

You can mix files and folders in one command. Each line of output reports OCR:, wrote:, or skip: for a file, and the run ends with a summary of written, skipped, and failed files. The command exits with a non-zero status if any file failed.

Output names

Each output keeps the source’s file name with a .txt extension: chapter-a.pdf becomes txt/chapter-a.txt. Batch output is always plain UTF-8 text. For Markdown or JSONL, use lexo ocr on single documents.

Resuming and overwriting

If a .txt with the target name already exists, LEXO skips that PDF. An interrupted batch can simply be run again: finished files are skipped and the rest are processed.

To replace existing files:

lexo ocr-batch ./pdfs --out-dir ./txt --overwrite

Embedded text in a batch

Batch OCR reads every page visually by default, because batches are usually scans and some scans carry poor hidden text layers. If your PDFs have good embedded text you want to keep, pass:

lexo ocr-batch ./pdfs --out-dir ./txt --no-force-ocr

The desktop dialog has the same choice: OCR every page, even if it has embedded text.

Tips for large batches

• Crop and split pages before batching if every PDF shares the same margins; batch OCR doesn’t crop.
• OCR speed depends on your network and Google Drive. LEXO works on a few pages at once and retries pages that fail transiently.
• Keep batches to documents you’re allowed to upload to your Google account.