一个独立的模型上下文协议（MCP）服务器，将AI编码助手与UCSF ChimeraX连接，用于分子可视化和结构生物学工作流程。

关于

ChimeraX MCP 弥合了对话式 AI 与分子可视化之间的鸿沟。你无需手动输入 ChimeraX 命令，而是用自然语言描述你想要的内容——开放结构、运行分析、生成发表数据——MCP 服务器通过其 REST API 将你的意图转化为精确的 ChimeraX 操作。

服务器自动处理所有事务：发现或启动 ChimeraX 实例、管理跨多个端口的会话、格式化结果，以及在出现问题时提供上下文错误提示。

关键能力：

• 119 专门工具涵盖完整的结构生物学工作流程
• 自动发射 — ChimeraX 在需要时自动启动，无需手动设置
• 多会话 — 同时处理多个ChimeraX实例
• 输入验证 — 原子规格验证及常见错误提示
• 可配置 — 端口、路径、超时和调试日志的环境变量

灵感来自RBVI官方ChimeraX MCP桥接器。

ChimeraX MCP Server

A standalone Model Context Protocol (MCP) server that connects AI coding assistants to UCSF ChimeraX for molecular visualization and structural biology workflows.

About

ChimeraX MCP bridges the gap between conversational AI and molecular visualization. Instead of manually typing ChimeraX commands, you describe what you want in natural language — open structures, run analyses, generate publication figures — and the MCP server translates your intent into precise ChimeraX operations via its REST API.

The server handles everything automatically: discovering or launching ChimeraX instances, managing sessions across multiple ports, formatting results, and providing contextual error hints when something goes wrong.

Key capabilities:

• 119 specialized tools covering the full structural biology workflow
• Auto-launch — ChimeraX starts automatically when needed, no manual setup
• Multi-session — work with multiple ChimeraX instances simultaneously
• Input validation — atomspec validation and error hints for common mistakes
• Configurable — environment variables for ports, paths, timeouts, and debug logging

AI Assistant  -->  MCP (stdio)  -->  chimerax-mcp (Python)  -->  HTTP REST  -->  ChimeraX

Inspired by the RBVI official ChimeraX MCP bridge.

Prerequisites

• UCSF ChimeraX installed
• Python 3.11+
• A local MCP client such as Claude Code, Claude Desktop, Codex, Cursor, VS Code, Gemini CLI, or similar

Quick Install (Codex Desktop, Codex CLI, and more)

git clone https://github.com/dovas-net/chimeraX-mcp.git
cd chimeraX-mcp
./install.sh

That’s it. The script creates a Python venv in .venv/, installs the package, writes the MCP config for Codex (shared by Codex Desktop and the Codex CLI via ~/.codex/config.toml), and runs a doctor check.

After the script finishes, fully quit Codex Desktop (⌘Q — not just closing the window) and reopen it. The ChimeraX tools will appear on the next launch.

Install for a different client instead:

./install.sh claude-desktop
./install.sh cursor
./install.sh claude-code
./install.sh copilot
./install.sh gemini
./install.sh cline
./install.sh continue
./install.sh windsurf

Windows users: install.sh is bash-only. Follow the Manual Installation steps below in PowerShell.

Manual Installation

If you’d rather drive each step yourself or you’re developing the server:

git clone https://github.com/dovas-net/chimeraX-mcp.git
cd chimeraX-mcp
python3 -m venv .venv && source .venv/bin/activate
pip install .                    # or `pip install -e ".[dev]"` for development
chimerax-mcp setup codex         # or any other client name