ABAQUS MCP Pro

Professional MCP server suite for Abaqus/CAE -- 100+ tools, dual transport, capsule tracking, physics contracts, and skills integration (Codex, Claude, etc.).

Architecture

MCP Client (Codex/Claude)
    |
    v  stdio
mcp server (server.py) <-- abaqus-mcp-pro package
    |
    v  TCP socket (localhost:48152)
Abaqus GUI plugin (gui_plugin.py / agent.py)
    |
    v  Abaqus Python API
Abaqus/CAE Kernel

The MCP server runs as a subprocess of the AI client. When the client invokes a tool (e.g. run_python), the server forwards the request over a TCP socket to a lightweight agent running inside Abaqus/CAE. The agent executes the code in the Abaqus Python kernel and returns results.

Related MCP server: COMSOL MCP Server

Features

• TCP socket bridge -- 10-50ms latency vs 100-200ms for file-based IPC

• AST error diagnostics -- auto-analyzes KeyError, AttributeError, NameError, TypeError and suggests fixes

• 22 MCP tools -- Full Abaqus automation: model query, job management, ODB inspection, KPI extraction, capsule tracking, physics contracts, report generation, viewport capture

• 13 MCP prompts -- Guided workflows for static, dynamic, modal, thermal, contact, fatigue, coupled, optimization, material, mesh, ODB extraction, and job debugging

• 74 MCP resources -- Session telemetry, skills knowledge base, CAE-Agent-Hub integration

• AST error diagnostics -- Auto-analyzes KeyError, AttributeError, NameError, TypeError and suggests fixes

• 3 CLI tools -- check, doctor, setup

• Dual transport -- TCP socket (primary) + file IPC (fallback)

• noGUI mode -- run Abaqus in batch mode without GUI

Installation

Prerequisites

• Python 3.10+

• Abaqus 2024+ (with Python 3.10)

• MCP-compatible AI client (Codex, Claude Desktop, etc.)

Install the package

pip install -e .

Install the Abaqus GUI plugin

abaqus-mcp-pro-setup

Or manually copy src/abaqus_mcp_pro/gui_plugin.py to your Abaqus plugins directory (typically ~/abaqus_plugins/).

Usage

1. Start Abaqus/CAE

Launch Abaqus/CAE normally. Then activate the plugin:

Plug-ins > ABAQUS MCP Pro > Start MCP Bridge

2. Configure your MCP client

Add to your MCP client configuration (e.g. mcp_config.json for Codex):

{
  "mcpServers": {
    "abaqus-mcp-pro": {
      "command": "abaqus-mcp-pro-server",
      "args": []
    }
  }
}

3. Use the tools

Once connected, the AI client can use any of the 22 tools:

noGUI Mode

Run Abaqus in batch mode with the TCP agent:

abaqus cae noGUI=scripts/start_abaqus_mcp_pro_agent.py

Or with the file IPC fallback:

abaqus cae noGUI=scripts/start_abaqus_mcp_pro_ipc.py

CLI Diagnostics

# Check connectivity to the running Abaqus bridge
abaqus-mcp-pro-check

# Full diagnostics
abaqus-mcp-pro-doctor

# Install/update GUI plugin
abaqus-mcp-pro-setup

Environment Variables

Examples

See the examples/ directory for Abaqus Python scripts:

• abaqus_cantilever_classic.py -- Cantilever beam static analysis

• abaqus_tensile_bar_classic.py -- Tensile bar with necking

• show_tensile_result_viewport.py -- Post-processing visualization

Project Structure

abaqus-mcp-pro/
+-- pyproject.toml           # Package metadata
+-- README.md                # This file
+-- src/abaqus_mcp_pro/
|   +-- __init__.py          # Package init
|   +-- server.py            # MCP stdio server (entry point, 67 lines)
|   +-- tools.py             # 22 MCP tools with register_tools()
|   +-- resources.py         # 3 MCP resources with register_resources()
|   +-- prompts.py           # 13 MCP prompts with register_prompts()
|   +-- skills.py            # 74 MCP skill resources with register_skill_resources()
|   +-- transport.py         # Socket + file IPC transport layer
|   +-- solver_diagnosis.py  # Solver Doctor: 40+ error patterns
|   +-- odb_lens.py          # ODB Lens: KPI extraction
|   +-- capsule.py           # Capsule: experiment state tracking & diff
|   +-- contracts.py         # Physics contracts validation
|   +-- report.py            # Simulation report generation
|   +-- agent.py             # Abaqus-side TCP socket agent (pure stdlib)
|   +-- gui_plugin.py        # Abaqus/CAE GUI plugin (AFX menu)
|   +-- file_ipc_plugin.py   # File-based IPC fallback plugin
|   +-- client.py            # TCP client for CLI tools
|   +-- protocol.py          # Shared line-delimited JSON protocol
|   +-- cli.py               # CLI: check, doctor, setup
+-- scripts/
|   +-- start_abaqus_mcp_pro_agent.py   # noGUI launcher (TCP)
|   +-- start_abaqus_mcp_pro_ipc.py     # noGUI launcher (file IPC)
|   +-- stop_mcp_agent.py           # Stop signal for file IPC
+-- examples/
|   +-- abaqus_cantilever_classic.py
|   +-- abaqus_tensile_bar_classic.py
|   +-- show_tensile_result_viewport.py
+-- tests/                    # Test directory

Credits

Built by integrating the best features from:

• Abaqus-Control-MCP -- TCP socket bridge, AST diagnostics

• CAE-Agent-Hub -- High-level tools, skills, and architecture

• abaqus-mcp -- File-based IPC transport

• Codex_MCP_Abaqus -- noGUI mode and examples