CarrotAI
@xingsandesu
About CarrotAI
CarrotAI is a cutting-edge AI agent application that delivers real-time streaming chat via Server-Sent Events (SSE) with built-in Model Control Protocol (MCP) integration. It supports concurrent connections to multiple SSE MCP servers and provides user interfaces in English, Chin
Overview
What is CarrotAI?
CarrotAI is a cutting-edge AI agent application that delivers real-time streaming chat via Server-Sent Events (SSE) and streamable HTTP with built-in Model Control Protocol (MCP) integration. It runs on a Flutter frontend + FastAPI backend and provides interfaces in English, Chinese, and Japanese for mobile, desktop, and web platforms.
How to use CarrotAI?
Clone the repository, set up environment variables (backend/.env and .env), start PostgreSQL, run the backend startup script, then deploy with Docker Compose. For manual setup, initialize the database and config, then run the backend server (python main.py). For the frontend, fetch Flutter dependencies, generate localization files, and run the app (flutter run). After launching, register/login via JWT, and interact with the AI agent in chat mode, optionally switching between SSE MCP servers in Settings.
Key features of CarrotAI
- Real-time AI chat powered by SSE and MCP adapters
- Concurrent connections to multiple SSE MCP servers
- Full localization in English, Chinese, and Japanese
- Deep thinking mode for complex, multi-step queries
- Secure JWT-based authentication and registration
- Responsive UI with light/dark mode and custom theming
- File upload to attach and parse documents in conversations
Use cases of CarrotAI
- Aggregate intelligent responses by calling multiple SSE MCP servers simultaneously
- Perform advanced analysis and deep reasoning for complex or multi-step queries
- Communicate with an AI agent in English, Chinese, or Japanese for multilingual support
- Upload files to provide richer context during conversations
FAQ from CarrotAI
What MCP servers does CarrotAI support?
CarrotAI supports SSE (Server-Sent Events) MCP servers defined in backend/config/mcp_servers.json. Custom adapters can be added via JSON files in backend/config/app/.
Which AI models are supported?
The supported LLM is DeepSeek, as defined in model_configs.json. Additional models can be configured by editing that file.
What platforms can CarrotAI run on?
The Flutter frontend runs on mobile (iOS/Android), desktop (Windows/macOS/Linux), and web (via WASM build).
Is CarrotAI free and open-source?
CarrotAI is licensed under the CarrotAI Open Source License. An online version is available at chat.jintongshu.com, and a SaaS client can be downloaded.
Can I add my own MCP servers?
Yes. Add SSE endpoints to mcp_servers.json or create custom adapter JSON files in backend/config/app/, then restart the backend to apply changes.
Basic information
More Agent Frameworks MCP clients
mcp-cli
wong2A CLI inspector for the Model Context Protocol
Where On-Device and Cloud LLMs Meet
hazyresearchBig & Small LLMs working together
MCP CLI - Model Context Protocol Command Line Interface
chrishayukUnified MCP Client Library
pietrozullomcp-use is the easiest way to interact with mcp servers with custom agents
🚀 MCPOmni Connect - Universal Gateway to MCP Servers
Abiorh001MCPOmni Connect is an agent execution runtime that connects to multiple MCP servers via stdio, SSE, or streamable HTTP. It supports chat, autonomous agents, and planner-based orchestration treating each MCP server as a tool agent to enable dynamic multi-agent workflows across LLM
Comments