Smarter Automation for Complex Business Forms
In day-to-day enterprise operations, information often needs to be repeatedly entered across ERP, OA, CRM, financial systems, supply chain platforms, and various industry-specific systems. Manual data entry not only consumes time but is also prone to omissions, errors, and redundant operations.
Traditional RPA can execute repetitive tasks based on predefined rules, but it often requires ongoing maintenance when facing page structure changes, non-standard controls, complex popups, and CAPTCHA scenarios. Fully AI-driven browser operations, on the other hand, may face challenges with execution stability, reproducibility, and rising invocation costs.
To address these real-world scenarios, SmartFilling v2 combines deterministic Playwright execution with on-demand AI fallback: operations that can be described through structured instructions are executed deterministically by automation scripts; when encountering complex popups, non-standard components, CAPTCHAs, or deterministic step failures, AI intervenes to handle the situation. This design balances execution stability and reproducibility with adaptability in complex scenarios.
Complete Pipeline: Recording, Orchestration, Execution, and Exception Handling
According to project documentation, SmartFilling currently supports recording and playback, script orchestration, browser operations, conditional logic, data extraction, file uploads, page screenshots, tab management, and CAPTCHA handling.
The project scripts offer three execution modes — sequential, loop, and AI-autonomous — which can be combined based on business process complexity. Developers can also exercise finer control over business flows through stage conditions, step conditions, error retries, fallback strategies, and AI fallback mechanisms.
For scenarios such as detail table entry, file uploads, complex page navigation, popup search, and page result detection, SmartFilling can organize execution logic through field variables, nested data, conditional branching, and flow jumps. The project also provides JSON Schema support, making it easier for developers to get field hints, format validation, and configuration assistance when editing scripts.
System Architecture
SmartFilling's system architecture consists of an automation execution engine, backend API, web frontend, background task execution service, test projects, and CAPTCHA recognition service. The project primarily uses .NET 8, Microsoft Playwright, OpenAI-compatible SDK, Jint, FastAPI, and ddddocr, providing a modular foundation for future feature expansion, component replacement, and scenario adaptation.
From "Providing Products" to "Opening Capabilities"
SmartFilling's public release on GitHub is not just the launch of a single project — it represents Workopilot's exploration of driving technology openness and ecosystem collaboration.
Workopilot is positioned as a digital employee platform for enterprise business processes, connecting existing systems, knowledge bases, models, and tools to enable digital employees to participate in business processing, task execution, and cross-system collaboration. The web operation and smart filling capabilities provided by SmartFilling serve as a crucial technical component for digital employees to connect with external business systems and execute repetitive data entry tasks.
By open-sourcing the project code and technical documentation, developers can gain a more intuitive understanding of SmartFilling's architecture and execution mechanisms, and conduct research and adaptation around different web systems, business processes, and industry scenarios. The project repository already includes contribution guidelines — developers can report issues or suggest features through GitHub Issues, or submit code following the project's contribution process.
Clear Licensing Boundaries
SmartFilling is released under the PolyForm Noncommercial License 1.0.0. Under this license:
- Personal learning, research, and non-profit organizations may use the software free of charge for non-commercial purposes
- Internal enterprise use, commercial product integration, and commercial service provision require appropriate commercial licensing
This licensing arrangement opens the code and promotes technical exchange while clearly defining the boundary between non-commercial use and commercial application.
Continuous Improvement of Smart Filling Capabilities
The public release of SmartFilling marks a starting point for Workopilot's technology openness. Going forward, Workopilot will continue to enhance SmartFilling's execution stability, script configuration experience, complex page adaptation, exception handling, and scenario expansion capabilities based on developer feedback and enterprise business needs — and explore its application in financial document entry, order processing, supply chain collaboration, project submissions, business data collection, and more.
Workopilot looks forward to engaging with more developers, software service providers, and enterprise users around smart filling, browser automation, and enterprise digital employees, working together to bring AI capabilities deeper and more reliably into enterprise business processes.
About Workopilot
Workopilot is a digital employee platform for enterprise business processes. Centered around people, processes, knowledge, devices, and systems, it provides an AI digital employee platform, intelligent hardware terminals, and industry solutions. The platform is committed to connecting existing business systems, knowledge bases, models, and tools, helping enterprises build configurable, operable, and manageable digital employees, driving the gradual intelligentization of repetitive business processing and cross-system collaboration.