1. Project Overview

Open-source enterprise IM + online customer service + knowledge base + ticketing system + AI Q&A. An open-source alternative to Feishu, DingTalk, and WeCom.

Weiyu - Conversation as a Service

An enterprise-grade, multi-tenant instant messaging solution featuring 8 free and open-source modules: Enterprise IM, Online Customer Service, Enterprise Knowledge Base / Help Documentation, Voice of the Customer, Ticketing System, AI Conversations, Workflow, and Project Management.

:::tip Weiyu is still in its early stages of rapid iteration. The documentation may lag behind development, so feature descriptions may not be accurate. Please refer to the latest released software version. :::

2. Open Source License

Licensed under AGPL-3.0

3. Screenshots

4. Feature Overview

Introduction

Enterprise IM

• LAN Instant Messaging
• Enterprise Member Management
• Chat History Monitoring
• ...

Omnichannel Customer Service

• Multi-Channel Integration
• Human Customer Support
• Customer Service Agent with AI capabilities, integrated with proprietary data to execute operations automatically
• ...

Knowledge base

• LLM Integration
• Custom Knowledge Base
• Function Calling
• Mcp
• ...

Ticketing System

• Ticket Management
• Ticket SLA Management
• Ticket Statistics and Reports
• ...

AI Large Language Models

• Ollama/DeepSeek/ZhipuAI/...
• Intelligent entity
• Workflow
• ...

Voice of the Customer

• Feedback
• Service Complaints
• Surveys
• ...

Project Dashboard

• To-Do Cards
• Daily/Weekly Reports
• Statistics
• ...

Workflow

• Custom Forms
• Custom Workflows
• Visualized Ticket Workflows
• ...

Social Groups

• Similar to Discord
• ...

Multi-Tenant

• Multi-tenant management
• Tenant Isolation
• Tenant Statistics
• ...

5. Tech Stack

Tech Stack

• springboot-3.x for Backend
• react for Web Frontend
• flutter for Mobile Client (ios&android)
• electron for Desktop Client (windows&mac&linux)

Quick Start with Docker

Method 1: Clone the project and start the docker compose containers. Ollama must be installed separately. The default model is qwen2.5:1.5b.

git clone weiyu && cd weiyu/deploy/docker && docker compose -p weiyu -f docker-compose.yaml up -d

Since the project uses the ollama qwen2.5:1.5b model by default, you need to pull the model in advance. Other models, such as deepseek-r1, can be configured in the configuration file.

ollama pull deepseek-r1:1.5b
ollama pull qwen2.5:1.5b

Method 2: Use docker compose ollama. Ollama will be installed by default, with qwen2.5:1.5b as the default model.

git clone weiyu && cd weiyu/deploy/docker && docker compose -p weiyu -f docker-compose-ollama.yaml up -d

Pull the ollama model using docker. Other models, such as deepseek-r1, can be configured in the configuration file.

docker exec ollama-bytedesk ollama pull deepseek-r1
docker exec ollama-bytedesk ollama pull qwen2.5:1.5b

If you do not need the AI-powered knowledge base Q&A feature, you can modify docker-compose.yaml or docker-compose-ollama.yaml to disable Ollama chat and embedding features to save resources.

# Disable Ollama chat
SPRING_AI_OLLAMA_CHAT_ENABLED: false
# Disable Ollama embedding
SPRING_AI_OLLAMA_EMBEDDING_ENABLED: false

Stop Containers

docker compose -p weiyu -f docker-compose.yaml stop

Modify the configuration; otherwise, uploaded images, files, and the knowledge base will not display properly.

• Modify the docker-compose.yaml or docker-compose-ollama.yaml file and update the following configuration items:

# Replace 127.0.0.1 with your server's IP address
BYTEDESK_UPLOAD_URL: [http://127.0.0.1:9003
BYTEDESK_KBASE_API_URL](http://127.0.0.1:9003
BYTEDESK_KBASE_API_URL): [http://127.0.0.1:9003](http://127.0.0.1:9003)

Method 3: BaoTa Panel

• Deployment Using BaoTa Panel

Method 4: Run from Source Code

• Run from Source Code