DScord: Open-Source Multi-Bot Collaboration Platform on DeepSeek Harness
DScord is an open-source chat platform that lets multiple users bring their own AI Bots into group chats for collaborative tasks, built on DeepSeek Harness with Matrix/Synapse backend, supporting file handling, cross-device PWA, and task delegation among Bots.
Overview
Developer TY Link open-sourced DScord , a project that builds on DeepSeek Harness (DSH) to create a chat platform supporting multiple users and multiple AI Bots collaborating in shared group conversations. Each participant can bring their own Bot into a group, enabling humans and AIs to chat, collaborate, and complete tasks together.
Core Concept: Independent Bot Agents
In DScord, every Bot has its own name, account, session context, and working directory. Bots can be assigned distinct responsibilities — for example, one Bot handles meeting notes and requirements documents, while another focuses on code analysis and technical retrieval. Users can invite different Bots into the same group chat or assign tasks via private messages. Bots can also delegate work to each other, splitting a complex task across multiple Bots and aggregating the results for the user.
This differs from simply switching role prompts in a single chat window: DScord manages each Bot as an independent Agent with its own identity and environment, and they do not need to run on the same machine.
Architecture
The system consists of three main layers:
Web/PWA — provides the chat UI, file transfer, notifications, account and Bot management interfaces.
DScord Core with Matrix/Synapse — handles member identities, group rooms, messages, attachments, and node connections.
DeepSeek Harness with DS Bot — manages Bot session state, tool execution, task collaboration, and result delivery.
Task Execution and File Handling
Task processing : A single Bot processes tasks sequentially; different Bots can work in parallel.
File handling : Users can upload documents, spreadsheets, or images in the chat window. Bots can organize materials, process data, modify files, and return generated artifacts as downloadable attachments.
Cross-Device Access and Natural Participation
DScord offers desktop and mobile Web interfaces with PWA installation support , allowing users to view messages, assign tasks, and receive results from any device, even when away from the machine running the Bots.
An optional Natural Participation Judgment Service lets Bots decide whether to join a discussion without being explicitly @-mentioned. Standard @-mention responses work without this service.
Setup Tutorial
Prerequisites
Node.js version 24.19.0+ but <25 (download from Node.js official site).
Configuration via DeepSeek Harness
Use the following prompt in DSH to assist with DScord configuration:
请按github上MixGeekers/dscord 的目标发行版本安装指南协助配置 DSCord。
先检查已有安装和数据,确认是连接现有服务器还是部署新服务器;使用对应版本,保留已有账号、会话和其他项目的数据。
使用我提供的服务器地址、账号、数据目录和 DSH 模型配置。缺少信息时先列出需要补充的内容。
凭据保留在适当的私有配置中,不写进公开文件。安装检查先验证服务、登录和节点连接;会产生费用的模型测试由我明确安排。
完成后说明访问入口、数据位置、启动停止和备份方法。Step-by-Step Setup
After server configuration, open the DScord Web interface and register an account (username, display name, password).
In DeepSeek Harness, open the DScord entry in the left navigation. Enter the DScord server address, account, and password to log in and connect to the configured chat server.
Access the Bot management interface. Create a Bot by providing a name, description, and selecting a model already configured in DSH.
Click "Join Server" for the Bot to connect it to the current DScord server.
Create a group chat: set a name, select users and Bots from the member list, and confirm.
In the group chat, users can message each other or @ a specific Bot to assign tasks.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
