AI Tools

Scivics Lab provides a suite of lightweight, Java/Quarkus-based tools for working with AI models and the Model Context Protocol (MCP).

LLM Chat UI

Tool	Description	GitHub
quarkus-chat-ui	Multi-provider web UI for LLMs — Claude Code CLI, OpenAI Codex CLI, vLLM, Ollama	quarkus-chat-ui

A unified browser-based chat interface that supports multiple LLM backends. Real-time SSE streaming, prompt queue, MCP server endpoint for agent-to-agent communication, and 10 color themes.

MCP Infrastructure

Tool	Description	GitHub
MCP Gateway	Name-based reverse proxy with caller identification and session metadata	quarkus-mcp-gateway
Emacs MCP Server	MCP server that controls Emacs via emacsclient	emacs-mcp-server

How They Fit Together

Claude Code / quarkus-chat-ui / Workflow Editor
        │  MCP
        ▼
quarkus-mcp-gateway  (:8888)
  ├── /mcp/chat-ui-28010  →  quarkus-chat-ui   (:28010)
  ├── /mcp/chat-ui-28020  →  quarkus-chat-ui   (:28020)
  └── /mcp/emacs          →  emacs-mcp-server  (:8092)

The gateway provides name-based routing (no need to remember port numbers), caller identification (each request carries metadata about who sent it), and a session metadata API for on-demand introspection.

Multiple quarkus-chat-ui instances can talk to each other via MCP. When Instance B receives a prompt with _caller pointing back to Instance A, it can reply by calling submitPrompt back on A. The conversation continues autonomously while humans watch from their browsers.

Multi-Agent Tutorials

Step-by-step guides for setting up agent-to-agent communication:

Tutorial	Description
Two Claude Code CLI Agents	Two instances both using Claude Code CLI
Claude Code CLI + Codex	Heterogeneous setup with Anthropic and OpenAI
Claude Code CLI + Local LLM	Using claw-code-local with Ollama

For three or more agents, the MCP Gateway provides name-based routing instead of direct peer-to-peer registration.