Summary

This project provides an intelligent customer service system for e-commerce platforms, specifically Pinduoduo. It aims to enhance customer support efficiency by offering AI-powered automated replies and a smart keyword-based system for transferring complex queries to human agents, benefiting merchants by streamlining operations and improving service quality.

How It Works

The system leverages Python's asyncio for asynchronous backend operations, integrating with various AI models (OpenAI, DeepSeek, Gemini, Kimi, Claude) via OpenAI-compatible APIs for intelligent response generation. Browser automation using Playwright facilitates interaction with e-commerce platforms. A core feature is its keyword-driven logic for identifying and routing complex customer issues to human agents, complemented by a knowledge base for custom and AI-generated replies. The architecture includes a PyQt6 frontend, SQLAlchemy/SQLite for data, and LanceDB for vector storage.

Quick Start & Requirements

• Primary Install/Run:

  git clone https://github.com/JC0v0/Customer-Agent.git
  cd Customer-Agent
  pip install uv uv venv
  uv sync
  uv run playwright install chrome
  python app.py
  

• Prerequisites: Python 3.11+, Windows 10/11 (recommended), stable network. Requires configuration of Pinduoduo merchant accounts, AI models, and reply templates.
• Links: GitHub Issues for feedback.

Highlighted Details

• Supports multiple LLM providers (OpenAI, DeepSeek, Gemini, Kimi, Claude) through a unified API interface.
• Tailored for Pinduoduo merchants with features like automated account login and status monitoring.
• Combines AI automation with intelligent keyword-based routing to human agents for comprehensive support.
• Features a desktop GUI built with PyQt6 and PyQt-Fluent-Widgets.
• Utilizes LanceDB for efficient vector-based knowledge retrieval.

Maintenance & Community

The project is primarily maintained by JC0v0. Community interaction is encouraged via GitHub Issues for feedback and suggestions. A technical communication channel is mentioned but not directly linked.

Licensing & Compatibility

The project is licensed under the permissive MIT License, allowing for broad compatibility with commercial use and integration into closed-source applications.

Limitations & Caveats

The system is recommended for Windows 10/11, and specific OS dependencies for Playwright might exist. No explicit mention of alpha/beta status, known bugs, or deprecations is present in the provided README.