MCP Materials Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@MCP Materials Server search for lithium cobalt oxide materials"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP Materials Server
A Model Context Protocol (MCP) server that provides AI assistants with access to materials science databases, starting with the Materials Project API.
Built with the MCP Python SDK using FastMCP.
Features
Tools (10)
Tool | Description |
| Search materials by chemical formula (e.g., "Fe2O3", "LiFePO4") |
| Retrieve crystal structure in CIF, POSCAR, or JSON format |
| Get comprehensive material properties (band gap, formation energy, etc.) |
| Side-by-side comparison of multiple materials |
| Find materials containing/excluding specific elements |
| Search by electronic band gap range (eV) |
| Find materials with similar crystal structures |
| Phase stability analysis for chemical systems |
| Mechanical properties (bulk/shear modulus, Debye temperature) |
| Find materials by mechanical property ranges |
Resources (2)
Resource | URI | Description |
Periodic Table |
| Element data with atomic numbers and masses |
Crystal Systems |
| The 7 crystal systems with symmetry constraints |
Prompts (3)
Prompt | Description |
| Comprehensive analysis workflow for a material ID |
| Search for battery electrode candidates |
| Compare phases in an alloy system |
Related MCP server: Materials Project Platform MCP Server
Installation
Prerequisites
Python 3.11 or higher
Materials Project API key (get one free)
Setup
# Clone or navigate to the project
cd mcp-materials-server
# Create virtual environment
python -m venv .venv
# Activate virtual environment
# On Windows:
.venv\Scripts\activate
# On macOS/Linux:
source .venv/bin/activate
# Install the package
pip install -e ".[dev]"Set API Key
# On Windows (PowerShell):
$env:MP_API_KEY = "your_api_key_here"
# On Windows (CMD):
set MP_API_KEY=your_api_key_here
# On macOS/Linux:
export MP_API_KEY="your_api_key_here"Usage
Run the Server
# Using the installed command
mcp-materials
# Or run directly
python -m mcp_materials.serverClaude Desktop Integration
Add to your Claude Desktop configuration file:
Location:
Windows:
%APPDATA%\Claude\claude_desktop_config.jsonmacOS:
~/Library/Application Support/Claude/claude_desktop_config.json
Configuration:
{
"mcpServers": {
"materials": {
"command": "python",
"args": ["-m", "mcp_materials.server"],
"cwd": "D:\\path\\to\\mcp-materials-server",
"env": {
"MP_API_KEY": "your_api_key_here"
}
}
}
}After adding the configuration, restart Claude Desktop.
Example Queries
Once connected to Claude, you can ask:
Basic Searches
"Search for lithium cobalt oxide materials"
"Find materials with formula Fe2O3"
"Search for materials containing Li, Fe, and O"
Property Lookups
"Get the properties of mp-149 (Silicon)"
"What is the band gap of mp-19017?"
"Get the crystal structure of mp-149 in CIF format"
Advanced Analysis
"Find materials with band gap between 1.5 and 2.5 eV"
"Get the elastic properties of silicon (mp-149)"
"Generate a phase diagram for the Li-Fe-O system"
"Compare the properties of LiCoO2 and LiFePO4"
"Find stiff materials with bulk modulus > 200 GPa"
Research Workflows
"Analyze material mp-149 comprehensively"
"Find potential Li-ion battery cathode materials"
"Compare phases in the Fe-Cr-Ni alloy system"
Development
Run Tests
# Run all tests
pytest
# Run with verbose output
pytest -v
# Run specific test class
pytest tests/test_server.py::TestToolFunctions -vLint Code
# Check for issues
ruff check src/
# Auto-format
ruff format src/Project Structure
mcp-materials-server/
├── src/
│ └── mcp_materials/
│ ├── __init__.py # Package version
│ └── server.py # MCP server (10 tools, 2 resources, 3 prompts)
├── tests/
│ ├── __init__.py
│ ├── conftest.py # Pytest configuration
│ └── test_server.py # Comprehensive test suite
├── pyproject.toml # Project configuration
├── claude_desktop_config.example.json
├── .gitignore
└── README.mdAPI Reference
Tool Details
search_materials(formula, max_results=10)
Search by chemical formula. Returns material IDs, band gaps, formation energies, and stability.
get_structure(material_id, format="cif")
Get crystal structure. Formats: cif, poscar, json.
get_properties(material_id)
Full property set: composition, symmetry, electronic, thermodynamic properties.
compare_materials(material_ids)
Compare list of materials side-by-side.
search_by_elements(elements, exclude_elements=None, max_results=10)
Find materials by element composition.
search_by_band_gap(min_gap=0, max_gap=10, direct_gap_only=False, max_results=10)
Search by band gap range in eV.
get_similar_structures(material_id, max_results=5)
Find materials with same space group.
get_phase_diagram(elements)
Build phase diagram for chemical system. Returns stable/unstable phases with decomposition products.
get_elastic_properties(material_id)
Mechanical properties: bulk modulus, shear modulus (Voigt/Reuss/VRH), Poisson ratio, Debye temperature.
search_by_elastic_properties(min_bulk_modulus=None, max_bulk_modulus=None, min_shear_modulus=None, max_shear_modulus=None, max_results=10)
Filter materials by mechanical properties.
Roadmap
Add AFLOW database integration
Add OQMD database support
Add electronic structure (DOS, band structure) tools
Add XRD pattern simulation
Add synthesis route suggestions
Add surface/interface properties
License
MIT
References
Author
Hesham Salama
Available Tools
10 toolscompare_materialsB
Compare properties of multiple materials side by side.
Args:
material_ids: List of Materials Project IDs (e.g., ["mp-149", "mp-66"])
Returns:
JSON table comparing key properties across materials
Inputs: material_ids (required). Returns: result
get_elastic_propertiesB
Get elastic and mechanical properties for a material.
Args:
material_id: Materials Project ID (e.g., "mp-149" for Silicon)
Returns:
JSON with elastic properties including bulk modulus, shear modulus,
Young's modulus, Poisson's ratio, and elastic tensor
Inputs: material_id (required). Returns: result
get_phase_diagramB
Get phase diagram data for a chemical system.
Args:
elements: List of elements defining the system (e.g., ["Li", "Fe", "O"] for Li-Fe-O system)
Returns:
JSON with phase diagram entries including stable phases, formation energies, and decomposition products
Inputs: elements (required). Returns: result
get_propertiesB
Get detailed properties for a specific material from Materials Project.
Args:
material_id: Materials Project ID (e.g., "mp-149" for Silicon)
Returns:
JSON with comprehensive material properties
Inputs: material_id (required). Returns: result
get_similar_structuresC
Find materials with similar crystal structures.
Args:
material_id: Materials Project ID to find similar structures for
max_results: Maximum number of similar structures (default: 5)
Returns:
JSON with structurally similar materials
Inputs: material_id (required), max_results. Returns: result
get_structureA
Get the crystal structure for a material from Materials Project.
Args:
material_id: Materials Project ID (e.g., "mp-149" for Silicon)
format: Output format - "cif", "poscar", or "json" (default: "cif")
Returns:
Crystal structure in the requested format
Inputs: material_id (required), format. Returns: result
search_by_band_gapB
Search for materials by band gap range.
Args:
min_gap: Minimum band gap in eV (default: 0)
max_gap: Maximum band gap in eV (default: 10)
direct_gap_only: Only return materials with direct band gaps
max_results: Maximum number of results (default: 10)
Returns:
JSON with materials in the specified band gap range
Inputs: min_gap, max_gap, direct_gap_only, max_results. Returns: result
search_by_elastic_propertiesB
Search for materials by elastic/mechanical properties.
Args:
min_bulk_modulus: Minimum bulk modulus in GPa
max_bulk_modulus: Maximum bulk modulus in GPa
min_shear_modulus: Minimum shear modulus in GPa
max_shear_modulus: Maximum shear modulus in GPa
max_results: Maximum number of results (default: 10)
Returns:
JSON with materials matching the elastic property criteria
Inputs: min_bulk_modulus, max_bulk_modulus, min_shear_modulus, max_shear_modulus, +1 more. Returns: result
search_by_elementsB
Search for materials containing specific elements.
Args:
elements: List of elements that must be present (e.g., ["Li", "Fe", "O"])
exclude_elements: Optional list of elements to exclude
max_results: Maximum number of results (default: 10)
Returns:
JSON with matching materials
Inputs: elements (required), exclude_elements, max_results. Returns: result
search_materialsB
Search for materials by chemical formula in the Materials Project database.
Args:
formula: Chemical formula (e.g., "Fe2O3", "LiFePO4", "Si")
max_results: Maximum number of results to return (default: 10)
Returns:
JSON with matching materials including material_id, formula, and key properties
Inputs: formula (required), max_results. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Each tool has a clearly distinct purpose with no significant overlap. For example, get_properties provides comprehensive details, while get_elastic_properties focuses specifically on mechanical properties, and get_structure returns crystal structure data. The search tools target different criteria (elements, band gap, elastic properties, formula), making tool selection straightforward.
All tools follow a consistent verb_noun naming pattern using snake_case. The pattern is uniform throughout: compare_materials, get_elastic_properties, get_phase_diagram, get_properties, get_similar_structures, get_structure, search_by_band_gap, search_by_elastic_properties, search_by_elements, and search_materials. This consistency makes the tool set predictable and easy to navigate.
With 10 tools, the server is well-scoped for materials science data retrieval. Each tool serves a specific function in querying and analyzing materials data, from basic searches to detailed property comparisons. The count aligns with typical MCP servers and covers the domain without being overwhelming or insufficient.
The tool set provides strong coverage for materials data retrieval, including search, get, and compare operations. However, there are minor gaps, such as no tools for updating or deleting data (though this may be intentional if read-only) and no specific tools for advanced analyses like thermodynamic stability or electronic structure beyond band gaps. Overall, it supports core workflows effectively.
Maintenance
Related MCP Connectors
Materials MCP — computed (DFT) materials structures & thermodynamic properties.
OQMD (Open Quantum Materials Database) MCP.
FTIR spectral search and material identification for AI agents via MCP.
Pay-per-use tool marketplace for AI agents. Search, price-check, and call APIs via MCP.
Related MCP Servers
- AlicenseAqualityDmaintenanceA tool for querying and analyzing materials data from the Materials Project database using natural language prompts, enabling materials scientists to explore properties, structures, and compositions of materials through conversational interfaces.31BSD 3-Clause
- AlicenseNot gradedqualityNot gradedmaintenanceThe Materials Project MCP Server is server that provides programmatic access to the Materials Project database. It enables LLMs, to search, analyze, and retrieve up to date materials science data.13MIT
- AlicenseNot gradedqualityCmaintenanceProvides computed (DFT) materials structures and thermodynamic properties through the Pipeworx MCP gateway, enabling AI agents to query materials data via natural language or tool calls.82 npmMIT
- FlicenseNot gradedqualityDmaintenanceEnables querying and retrieving materials data from the Materials Project via MCP tools, supporting search, lookup, comparison, and export.1-