An application developed on UniAI and inspired by KimiChat, featuring integration with multiple AI models.

中文说明

Introduction

We are inspired by Moonshot and its product, KimiChat. This led us to develop and open-source LeChat Pro, a document parsing and Q&A chat platform based on generative AI models and knowledge graphs. LeChat Pro supports parsing of a full suite of Office documents, PDFs, images, and more, presented in a chat format similar to ChatGPT.

Our project is simple and community-driven, relying on open-source contributions to enhance functionality. We welcome you to join us in upgrading LeChat!

While we lack the resources to train our own large language models, this grants us flexibility in model choice, allowing integration with any model. Currently, LeChat Pro supports:

• Moonshot
• OpenAI
• iFlyTek Spark
• Baidu WenXin Workshop
• Google Gemini
• ZhiPu AI
• Alibaba Tongyi Qianwen

You can connect additional models by contributing to our other open-source project, UniAI, which is the core library for LeChat's multi-model support.

If you prefer open-source models over commercial ones, we offer support for the GLM model. For deployment, please follow the instructions at https://github.com/uni-openai/GLM-API, then configure the backend environment variables to add the GLM_API address.

Experience it here:

👉 https://lechat.cas-ll.cn

Open-source backend repository:

👉 https://github.com/uni-openai/uniai-maas

UniAI Core Library:

👉 https://github.com/uni-openai/uniai

Mini Program Repository:

👉 https://github.com/uniai-lab/lechat-miniapp-v1

Features

• 🧠 Multi-model switching
• 📃 Complete Office file parsing
• 👀 Image recognition
• 🎨 Image generation
• 📈 Chart generation

ChatKG Knowledge Graph

We are developing an innovative feature for chat based on knowledge graphs—ChatKG.

Quick Start

Note: This project requires the UniAI backend framework. Deploy it here.

Before getting started, ensure you have Node.js installed. If not, download it here.

Once set up, navigate to the project's root directory and run the following commands to start:

npm install
npm run dev

Or

yarn
yarn dev

Upon success, you will typically see output similar to:

VITE v3.2.5 ready in 294 ms

➜ Local: http://localhost:5173/
➜ Network: use --host to expose

Hold Ctrl or Command and click the Local link to open the project in your browser. You can log in with a QR code or mobile verification code to start using the application.

If you plan to package the project for local deployment, see here.

Contributors

Weilong Yu

Youwei Huang

This project is licensed under the MIT License.