Read Text That Was Never Text
Use OCR to capture text from images, screenshots, PDFs, and scanned documents directly inside your automation.
Unlock Image-Based Content
OCR expands what you can scrape beyond HTML text. Many pages store important information inside images, screenshots, embedded graphics, and scanned documents. Without OCR, that content is invisible to ordinary selector-based extraction. With OCR, it becomes a normal part of your dataset.
The built-in OCR command in RTILA X reads text from images directly inside the automation. You can capture a product image label, extract a number from a chart, or pull data from a scanned PDF. The result behaves like ordinary text data that you can export to CSV, Google Sheets, or a database.
When You Need OCR
OCR is valuable when text is not available in the page DOM:
- Product specifications stored in image labels
- Screenshots of dashboards or reports
- Scanned invoices and receipts
- PDFs that are not text-searchable
- CAPTCHA result pages that display text visually
- Memes, infographics, or charts with embedded text
Using OCR expands the range of sites and documents you can process. For some workflows, OCR is the only practical way to capture the data.
Supported Languages
RTILA X OCR supports text in many languages, which makes it suitable for international workflows. You can process English content alongside documents in French, German, Spanish, Arabic, Chinese, Japanese, and others.
The language setting can be configured per OCR command, so a single automation can handle mixed-language pages. This is especially useful for marketplaces, global directories, and research projects that span multiple regions.
How OCR Works in RTILA X
- The automation captures an image or screenshot.
- The OCR command analyzes the image and detects text regions.
- Recognized text is cleaned and normalized.
- The text is stored as a dataset field for the current row.
Because OCR runs locally, the image never leaves your machine unless you configure an external OCR integration. That keeps document images, screenshots, and other sensitive content under your control.
Combining OCR with Data Extraction
OCR does not replace standard selectors; it fills the gaps. A common workflow uses regular CSS extraction for page titles, prices, and links, then uses OCR for a small portion of the page that is rendered as an image. The two types of data combine into a single dataset row.
This mixed approach gives you complete coverage without trying to use OCR for everything. OCR is powerful, but it is slower and less exact than direct text extraction. Using it only where needed keeps automation speed high while still capturing difficult content.
Accuracy Tips
- Use high-quality screenshots or images when possible.
- Crop the image to the relevant text region before running OCR.
- Apply image preprocessing when available, such as increasing contrast.
Use It Alongside Structured Selectors
Combine OCR with normal selectors for complete data coverage. OCR expands what you can scrape beyond HTML text and makes image-based content part of your normal workflow.