Indexing and OCR

Uploading a book to Theca is the first step. Before you can search it or generate a summary from it, the book must go through an indexing process. This article explains what happens during that process and how to read the statuses the app shows.

Why processing is necessary

Theca's search does not work on the raw file. The system extracts the text, splits it into fragments, and builds a representation that makes it possible to find relevant passages even when you do not remember the exact words. Without that processing, the book can only be read and downloaded.

What the app shows in the meantime

Each book displays a status indicator on its card and in the detail panel:

• Pending — the book has not entered the processing queue yet.
• Queued — the book is waiting for its turn.
• Processing — indexing is in progress. The indicator may show a progress count (for example, "Processing 4/10").
• Processed — the book is ready for search and summaries.
• Error — retry — something failed. You can try again with the Retry button in the detail panel.

To queue a book manually, open its detail panel and click Process. To queue all pending books at once, use the Process all button in the toolbar.

Scanned PDFs: the special case

A PDF can consist entirely of page images with no text layer — what is known as a scan. Theca detects this automatically and shows the status Scanned, no text.

A book in that state can be read and downloaded, but it cannot be indexed directly because there is no text to extract. To make it searchable, you need to use optical character recognition (OCR):

1. Open the book's detail panel.
2. Click Extract text (OCR).

OCR analyzes each page as an image and extracts the text. This process takes significantly longer than regular indexing, especially for long books. While it runs, the status changes to Extracting…

OCR has a cooldown period: if you just launched it and it stops, you will need to wait a few minutes before you can start it again.

Capability badges

Once a book is processed, it may show additional badges indicating what kind of content was recognized:

• OCR — text was extracted via optical recognition (the book was a scan).
• Formulas — the book contains recognized math or chemistry notation.

These badges help you know what to expect when searching or summarizing the book.

What to do if a book shows an error

If the status shows Error — retry, open the detail panel and click Retry. If the error persists after several attempts, check that the file is not corrupted by downloading it and opening it on your device.

Frequently asked questions

How long does regular indexing take? It depends on the book's size and system load. For books of typical length it is usually a matter of minutes.

What about OCR? OCR takes significantly longer than regular indexing because it analyzes each page as an image. Long books can take several minutes.

Can I read a book while it is being processed? Yes. You can open it in the reader at any time, regardless of its indexing status.

The book has been "Queued" or "Processing" for a long time — is that normal? During periods of high activity the system may take longer. If the status does not advance after an extended period, try clicking Retry from the detail panel.

Next step

• Search your library — how search works once your books are processed.
• Summaries — generate cited summaries from your indexed books.
• Your library — go back to the library basics.