BotSharp UI

Positioning#

BotSharp UI is the official frontend project of the BotSharp multi-Agent framework under the SciSharp organization, serving as a centralized management frontend for AI Agents. It addresses the lack of a unified visual operation plane when enterprises build multi-Agent systems using the BotSharp backend framework. This repository contains only the frontend interface; all core capabilities such as LLM invocation, RAG retrieval, MCP tool integration, and multi-channel message routing are provided by the BotSharp backend (C# / .NET Core).

Core Capabilities#

Agent Building & Management#

• Node-based visual building: Node-based interaction for rapid AI assistant creation
• Agent lifecycle management: Create, configure, and manage existing Agent instances

Conversation & Real-time Interaction#

• Conversation session management: Manage conversation sessions and history
• WebSocket real-time communication: Realtime Session support with low-latency streaming
• Reasoning process visualization: Display Agent thinking/reasoning processes (refine thinking content)

Other Features#

• Built-in advanced search functionality
• Native Chinese internationalization support
• Containerized deployment support (Dockerfile)
• Multi-environment configuration (.env, .env.local, .env.production)
• Native Azure Static Web Apps deployment support

Frontend Architecture#

Modern SPA application based on SvelteKit v2 with Vite as the build tool. Technology composition: Svelte (48.4%), SCSS (41.8%), JavaScript (9.8%). The frontend is fully decoupled from the backend, communicating via RESTful API (standard requests) and WebSocket (real-time sessions), with build output as pure static files for independent deployment.

Backend Dependency (BotSharp)#

The BotSharp backend uses C# / .NET Core with a plugin-based + Pipeline streaming execution design where plugins are fully decoupled. Core modules include Plugin Loader, Hooking, Authentication, Agent Profile, Conversation & State, Routing & Planning, Templating, File Repository, Caching, Rich Content, and LLM Provider. Supported LLMs include OpenAI (GPT-3.5/4o/o1), Google Gemini 2, Anthropic Claude, DeepSeek V3, LLaMA 3, HuggingFace, etc. Storage supports MongoDB, LiteDB, and Tencent COS. Built-in capabilities include RAG, vector search, multi-Agent routing and planning, and MCP tool integration. Message channels cover Facebook Messenger, Slack, Telegram, etc.

Installation & Deployment#

Prerequisite: BotSharp backend service must be running first (default address http://localhost:5015/)

git clone https://github.com/SciSharp/BotSharp-UI
cd BotSharp-UI
npm install
npm run dev

Production Build:

npm run build
npm run preview

Azure Deployment:

npm run build -- --mode production
npm install -g @azure/static-web-apps-cli
swa deploy ./build/ --env production --deployment-token {token}

Configuration System#

• Root .env for default config, .env.local to override defaults (not version-controlled), .env.production for production
• Switch environment config files via --mode CLI parameter
• Specify BotSharp backend API address through environment variables

Use Cases#

• Enterprise internal AI Agent building and management platform
• Visual debugging and testing of multi-Agent collaboration systems
• LLM Agent conversation auditing and reasoning chain visualization
• Unified entry point for RAG knowledge base management and MCP tool management scenarios when paired with the backend

Pending Confirmation#

• Specific .env configuration fields not fully documented; backend API address environment variable name pending confirmation
• Latest Release tagged as r2.0-agent-utility, not using standard SemVer
• Online Demo URL not provided in README
• WebSocket connection path and message protocol format not documented
• Technical implementation details of Agent reasoning process visualization unclear