About pymol-cursor-mcp
pymol-cursor-mcp is an MCP server published by truong128 in the Files category: control PyMOL from Cursor via natural language. Enables describing molecular structures and styles in plain language to execute PyMOL commands through XML-RPC.
The repository has 1 stars and 0 forks, with the last commit 5 months ago. Recent commit activity is the strongest signal that reported issues will get attention.
Install
git clone https://github.com/truong128/pymol-cursor-mcpThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about pymol-cursor-mcp
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
pymol-cursor-mcp - Truong Nguyen
Control PyMOL from Cursor via the Model Context Protocol (MCP). Describe structures and styles in plain language; the assistant calls PyMOL through XML‑RPC (run_command, run_python, pymol_get).
This project packages and documents a Cursor-first workflow. The same MCP server works with Claude Code if you prefer that CLI.
Upstream inspiration: nagarh/pymol-claude-code (PyMOL + MCP for Claude Code). This repo extends that idea with portable Cursor config, checks, scripts, and install docs.
Architecture
Same diagram style as upstream flow.svg, updated for Cursor (cyan node). Uses SVG/CSS animation (animateMotion, keyframes). If your Markdown preview looks static, open flow.svg directly on GitHub for the full motion.
What you get
| Piece | Purpose |
|---|---|
pymol_mcp_server.py |
MCP server (stdio) → PyMOL XML‑RPC |
.cursor/mcp.json |
Registers the pymol MCP server using ${workspaceFolder} |
.cursor/rules/*.mdc |
Agent hints for PyMOL tool usage |
scripts/start_pymol_for_mcp.sh |
Launch PyMOL with -R (macOS/Linux) |
verify_stack.py |
Sanity check: venv + MCP import + RPC to PyMOL |
examples/restore_pocket_labels.pml |
Example PyMOL script (labels / pocket) |
flow.svg |
Animated architecture diagram (same style as upstream repo) |
Quick install
git clone https://github.com/truong128/pymol-cursor-mcp.git
cd pymol-cursor-mcp
python3 -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -U pip
pip install -r requirements.txt
deactivate
Install PyMOL (conda‑forge recommended), then start RPC:
conda activate pymol_mcp # your env with pymol-open-source
pymol -R
Or: ./scripts/start_pymol_for_mcp.sh (after chmod +x).
In Cursor: Open Folder → this repo → Settings → MCP → confirm pymol connected → restart Cursor if needed.
python3 verify_stack.py
Should print: OK: MCP import + PyMOL RPC at http://localhost:9123
Full step-by-step: INSTALL.md
Daily use
- Terminal:
pymol -Ror VS Code/Cursor task: “PyMOL: start with XML-RPC”. - Cursor: Chat / Agent with this folder open.
- Prompt: e.g. “Fetch 1hvr, cartoon on protein, sticks on ligand.”
Summary of changes vs “Claude Code only” upstream
| Topic | Here |
|---|---|
| IDE | Cursor primary; .cursor/mcp.json, rules, tasks |
| RPC URL | PYMOL_RPC_URL env (default http://localhost:9123) |
| Paths | ${workspaceFolder} — clone anywhere |
| Windows | .cursor/mcp.json.windows.example |
| Verify | verify_stack.py |
| PyMOL start | scripts/start_pymol_for_mcp.sh + Run Task |
Troubleshooting
- MCP disconnected: see Cursor Output → MCP Logs; recreate
venv; reopen repo root as workspace. - PyMOL errors / missing
.dylib: install PyMOL from conda-forge into a dedicated env (see INSTALL.md). - Remote/HPC: run MCP where your editor runs; tunnel RPC (
PYMOL_RPC_URL, SSH-R) — details in INSTALL.md.
Credits
- PyMOL MCP idea and original server: pymol-claude-code — Hemant Nagar.
- Packaging and Cursor-focused docs: truong128. - Truong Nguyen
License
MIT — see LICENSE.
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.