pmh1314520/WebRPA

WebRPA – A Visual RPA Platform with Built‑in AI Assistant

What it is – WebRPA is an open‑source, Windows‑only robotic‑process‑automation (RPA) suite that lets users build web, desktop and Android automation flows by dragging and connecting pre‑made modules. No programming is required; the editor generates the underlying Python/Node.js code for you. A bundled “AI little helper” can understand a natural‑language description of a task, auto‑assemble a workflow, and even diagnose and repair failures.

Why it matters – RPA is a fast‑growing way to automate repetitive business tasks. WebRPA goes a step further by providing:

• Zero‑code visual authoring – a React‑based canvas powered by React Flow.
• All‑in‑one runtime – the download includes a self‑contained Python 3.13 interpreter, Node 20, Playwright, FFmpeg, Pandoc, OCR libraries, etc., so the tool works offline without extra installs.
• AI‑driven workflow creation & self‑healing – compatible with any OpenAI‑style API (OpenAI, Deepseek, Ollama, etc.) and can automatically rebuild a broken flow.
• Modular feature packs – heavy‑weight capabilities (OCR, speech, media processing, mobile automation, QQ bots, …) are shipped as optional zip packs that can be added or removed on demand, keeping the core download lightweight.

Core capabilities (571 built‑in modules)

Category	Example actions
Web automation	Page navigation, element click/type, data scraping, anti‑detection via DrissionPage, network interception
Desktop automation	Mouse/keyboard simulation, image‑based clicking, screen capture, Windows API calls
Mobile automation	Android tap/ swipe / input, screen mirroring, app install/uninstall
Flow control	Conditions, loops, sub‑flows, timers, Webhook/ hotkey/file/email triggers
Data handling	Variable math, text ops, list/dict, CSV/Excel (openpyxl), DB (MySQL/PostgreSQL/Redis)
AI services	Chat, vision, extraction, classification, summarisation, translation, routing, OCR (EasyOCR, ddddocr), speech‑to‑text, face recognition, image generation
Media	Image edit, watermark, video/audio conversion (FFmpeg), PDF merge/split, document conversion (Pandoc)
Integration	HTTP requests, 17 messaging channels, QQ/WeChat bots, SSH, SAP, LAN sharing
Utilities	Encryption, color/time conversion, script execution (JS/Python), Allure test reports

How you run it

1. Download the latest 7z package from Releases.
2. Extract and double‑click WebRPA启动器.exe.
3. The launcher reads WebRPAConfig.json (host/port settings) and starts a FastAPI backend (default http://localhost:5241) and a React frontend (default http://localhost:5921).
4. Open the browser window, drag modules onto the canvas, connect them, set parameters, and hit Run.
5. Optional: install additional feature packs via the “More → 功能模块包” menu; the editor will prompt you if a required pack is missing.

Technical stack

• Frontend – React 19 + TypeScript 5, Vite, Tailwind 4, Radix UI, shadcn/ui, Zustand, Monaco editor, Framer Motion, Mermaid for flow diagrams.
• Backend – Python 3.13, FastAPI, Uvicorn, Pydantic 2, Socket.IO, Playwright (browser automation), Polars (data frames), httpx, openpyxl, PyMySQL, etc.
• AI integration – OpenAI‑compatible API client; supports multiple providers (OpenAI, Deepseek, Ollama, etc.).
• Media & vision – FFmpeg, Pillow, OpenCV, EasyOCR, ddddocr, SpeechRecognition, face_recognition.
• System automation – PyAutoGUI, pynput, pywin32, mss, mitmproxy.
• Marketplace service – Node 20 + Express + SQLite, managed by PM2.

Who it’s for

• Business users, QA testers, data analysts, and developers who need to automate repetitive web/desktop/mobile tasks but lack coding skills.
• Teams that require an offline, self‑contained solution (e.g., secure intranet environments).
• Organizations that want a free, AGPL‑licensed platform for internal use, with the option to purchase a commercial license for closed‑source deployment and support.

License & commercial model

• Open‑source – AGPL‑3.0 for personal, educational, and non‑commercial use. Any public network service must also be open‑sourced.
• Commercial – Paid private license permits closed‑source commercial deployment and includes technical support.

Bottom line – WebRPA bundles a full RPA engine, a modern visual editor, and an AI assistant into a single offline package. It lets non‑programmers create sophisticated automation flows across web, desktop, and mobile platforms, while still offering extensibility for developers who want to add custom modules or integrate their own AI services.