OCR, TOTP, LLM and office documents#

Text recognition#

Local OCR uses Tesseract 5 or later without a license-based recognition count limit. Blocks process images, pages, windows, elements and screen regions. Built-in Microsoft OCR is available on compatible workstations. The module, language and recognition area are selected before execution.

One-time passwords#

The TOTP block accepts a secret key and returns a time-based one-time password. The Robot uses it in the configured authentication step. The key is stored as a protected Asset and excluded from logs. See the TOTP reference.

Finding an element by description#

An LLM identifies a page or application element from a textual description and creates a selector for the block. The User checks the identified element before use. The selector is saved in the workflow and used by predefined interaction blocks. Model settings, page access and search errors are considered when configuring the Process.

Office files and macros#

The platform works with Excel, Word, PowerPoint presentations and open formats. Reading, writing and conversion work without an installed office suite. Excel supports filtering, sorting, sharing and row or column management. Word supports tables, formatting, headers, footers, bookmarks and macros.

The workflow specifies the file and processing mode. Executing an existing macro is a separate operation, with its result and the selected office environment’s compatibility checked. An R7-Office plugin extends document blocks. See the R7 block reference.

Workstation dependencies and application permissions must match the blocks used. Opening, processing and saving a document are separate Process steps.