By peter_lzh Xianyu Auto-Reply: An Open-Source Multi-Account Management System for Goofish Sellers
Managing multiple accounts on Xianyu (Goofish), China’s largest second-hand marketplace, usually means juggling endless chats and repetitive tasks. Xianyu Auto-Reply is an open-source management system built with FastAPI, SQLite, and Playwright that automates the whole workflow — multi-user and multi-account management, keyword replies, AI-powered auto-reply, automatic order fulfillment, product management, log monitoring, and one-click Docker deployment.
⚠️ Important: the project is released under the AGPL-3.0 license and is intended for learning and research purposes only. Do not use it for any illegal or non-compliant scenarios, and read the copyright notice carefully before use.
Core Features
- Multi-user system: registration and login, email verification, image captcha, permission control, and per-user data isolation.
- Multi-account management: each user can manage multiple Xianyu accounts with independent start/stop, status monitoring, and cookie maintenance.
- Smart replies: keyword replies, default replies, product-specific replies, image keywords, and AI auto-reply.
- Automatic fulfillment: supports text, batch data, API, and image-based delivery, with duplicate-processing protection.
- Product management: automatically collects product information, with product details, spec configuration, and data deduplication.
- Operations tooling: order management, notification channels, message alerts, and online customer service for back-office teams.
- Monitoring and maintenance: real-time logs, health checks, security statistics, system statistics, and log file rotation.
- Containerized deployment: run locally, via Docker Compose, or with multi-architecture builds.
Quick Start
Docker Compose (recommended)
Default endpoints after startup:
- Web admin interface:
http://localhost:9000 - API documentation:
http://localhost:9000/docs - Health check:
http://localhost:9000/health
Local run
Default address: http://localhost:8090
For more deployment options — Windows scripts, China-optimized Docker Compose configuration, multi-architecture builds, and access addresses — see the project’s deployment guide.
AI Reply Configuration
AI replies use a unified model_name / api_key / base_url / api_type configuration pattern, supporting OpenAI-compatible endpoints, OpenAI Responses, DashScope, Gemini, Anthropic, and Azure OpenAI.
Third-party OpenAI-compatible services can be plugged in via a custom base_url; see the configuration docs for details and instructions on adding new providers.
Technical Architecture
- Backend: FastAPI + Uvicorn + Python 3.11+ async programming
- Database: SQLite 3 with multi-user data isolation and automatic migrations
- Frontend: Bootstrap 5 + Vanilla JavaScript + Chart.js
- Protocols: REST API + WebSocket + SSE
- Automation: Playwright + DrissionPage
- Deployment: Docker + Docker Compose + optional Nginx
- Logging: Loguru with file rotation and real-time collection
Monitoring and Maintenance
- Real-time logs: view live system logs from the web interface.
- Log files: split by date under the
logs/directory. - Log levels: DEBUG, INFO, WARNING, and ERROR.
- Health check: hit
/healthto verify service status.