EMILY.RPA Documentation

The EMILY.RPA Desktop Automation module uses computer vision to analyze the screen, locate coordinates of reference patterns, and automate desktop applications by controlling the mouse and keyboard. To ensure reliability, the deployment environment's screen settings (resolution, color depth, text scaling) must match those used during the training phase. It is recommended to design workflows directly on the deployment machine and avoid manual computer operation during execution to prevent interference.

The File System Agent allows users to describe a task goal in natural language, which a Large Language Model (LLM) then executes by autonomously operating on a specified folder. Supported operations include reading, writing, and deleting files. To reduce failure rates and token consumption, it is recommended to break complex goals into smaller, manageable tasks. Users can use the 'Ask EMILY' feature to generate optimized prompts for the LLM.

The EMILY.RPA PowerShell module allows users to execute PowerShell scripts to automate local system tasks and control installed Microsoft Office applications. It provides both No-Code and Low-Code interfaces.

Remote Trigger allows EMILY.RPA to act as a process execution server. This enables users to call and execute specific skills from other devices and receive the results. Multiple EMILY.RPA instances can be linked to provide load balancing and high availability. Note that this feature is only available for Enterprise Management groups.

EMILY.RPA is a cross-platform Robotic Process Automation (RPA) software developed by BILINK CORP. It is designed to operate the user interfaces of web pages and desktop software, enabling automation for systems that lack available APIs or SDKs.

The PDF Form Analysis module is used to analyze PDF files and extract text information. It is highly recommended to use files that comply with the PDF/A standard. The module operates in two modes: 'VIEW' for document preview and visual coordinate generation, and 'TRAIN' for writing extraction code.

The Browser Use Agent module allows users to complete web-based tasks by describing a goal in natural language. An LLM autonomously operates the browser to achieve that goal. To improve success rates and reduce token consumption, it is recommended to break complex goals into multiple smaller tasks. Users can use the 'Ask EMILY' feature to generate optimized prompts for the LLM.

The EMILY.RPA Remote Desktop Automation module allows users to automate remote Windows applications by connecting via the Remote Desktop Protocol (RDP). It uses computer vision to analyze the screen and identify coordinates of reference patterns to operate the remote mouse and keyboard. The module also supports reading/writing to the remote clipboard and uploading/downloading files (which requires SSHD to be enabled on the remote host).

The PII Redaction module protects sensitive data by replacing PII (IDs, phone numbers, emails, etc.) with placeholders locally. This allows data to be sent to an LLM safely. After receiving the LLM's response, the placeholders are restored to their original values using a local mapping file. The process is fully asynchronous and ensures plain-text PII does not leave the local environment.

The EMILY.RPA Group feature allows administrators to centrally manage multiple accounts, skills, and licenses. It provides a centralized dashboard for auditing usage records sorted by time and managing organizational resources.

The Agentic Workflow module enables the automation of tasks described in natural language. A Large Language Model (LLM) autonomously plans the necessary steps to achieve the goal. The agent has the capability to operate a web browser, access files within the working folder, send emails via bound accounts, and write code. To improve success rates and reduce token consumption, it is recommended to break large goals into smaller, manageable tasks. Users can use the 'Ask EMILY' feature to generate optimized prompts for the LLM.

The EMILY.RPA Azure OCR module uses Azure cloud services to parse text and table information from image-based files. This module requires version 2.2.20230211 or higher.