ggui
Active·★ 42·Apache-2.0·Updated 2026-10-21
ggui is an open protocol and reference runtime that allows AI agents to generate ephemeral, interactive user interfaces over MCP without writing frontend code. It can be self-hosted and paired with any MCP-aware agent runtime such as Claude Desktop, Cursor, or ChatGPT desktop.
#agent-framework#agent-ui#agents#ai#generative-ui#llm#mcp#mcp-servers
01
Features
01Generates ephemeral, interactive UIs over MCP without frontend code
02Provides MCP tools: ggui_push, ggui_update, ggui_handshake and blueprint tools
03Supports embedding via <McpAppIframe> or direct iframe
04Can be self-hosted with zero cloud dependency
05Works with any MCP-aware agent runtime (Claude, OpenAI, Gemini, etc.)
02
Compatibility
Claude Desktop
Claude Desktop
Verified via docs
Claude Code
Claude Code
Verified via docs
Cursor
Cursor
Verified via docs
ChatGPT Desktop
ChatGPT Desktop
Verified via docs
Goose
Goose
Verified via docs
03
Quick start
1
$ npm create ggui-server@latest my-app
2
$ cd my-app
3
$ npm install
04
Use cases
↳AI agent pushes dynamic UIs to users' browsers in real-time
↳Embedding ggui sessions inside existing applications for agent-user interaction
↳Self-hosting the protocol or using a managed provider like Guuey
05
Alternatives
mcp-gateway★ 84
Universal MCP Gateway - Single-port multiplexing with Meta-MCP for ~95% context token savings
MCP-Plugin-dotnet★ 20
.NET MCP bridge: expose app methods/data as MCP tools, prompts, and resources via an in-app plugin + lightweight server (SignalR; stdio/http).
blockrun-mcp★ 391
BlockRun MCP Server — Access 41+ AI models in Claude Code with zero API keys. Pay-per-request via x402 micropayments.
MediaWiki-MCP-Server★ 138
Model Context Protocol (MCP) Server to connect your AI with any MediaWiki
Comments
Log in to leave a comment
No comments yet. Be the first!