ChemMCP - Chemical Molecule Computation Platform

ChemMCP is a suite of chemical molecule computation tools based on the Model Context Protocol (MCP), providing 19 professional tools for handling chemical molecules.

🚀 Quick Start

Environment Requirements

• Python 3.11+ (Python 3.11 or 3.12 recommended)
• Conda environment manager

Installation Steps

1. Create a conda environment:
   bash
   conda create -n chemmcp python=3.11
   conda activate chemmcp

2. Install dependencies:
   bash
   pip install -r requirements.txt

3. Install ChemMCP:
   bash
   pip install -e .

Installation Options

ChemMCP offers flexible installation options, allowing you to choose according to your needs:

Option 1: Basic Installation (Recommended)

bash

Installs core dependencies, supporting 17 tools

pip install -r requirements.txt

Supported Tools: All tools except MoleculeCaptioner and MoleculeGenerator

Option 2: Full Installation (Includes AI Tools)

bash

Method A: Using optional dependency groups

pip install -e ".[ai]"

Method B: Manually installing AI dependencies

pip install -r requirements.txt
pip install transformers>=4.20.0 torch>=1.9.0

Supported Tools: All 19 tools, including AI-driven molecular captioning and generation capabilities

Option 3: Development Environment

bash

Installs all dependencies and development tools

pip install -e ".[all,dev]"

Setting Up Environment Variables

Set the appropriate API keys based on the tools being used:

bash

Required for WebSearch tool

export TAVILY_API_KEY="your_tavily_api_key"

Required for LLM tools

export LLM_MODEL_NAME="openai/gpt-4o"
export OPENAI_API_KEY="your_openai_api_key"

Optional for ChemSpace tool (used as fallback if PubChem lookup fails)

export CHEMSPACE_API_KEY="your_chemspace_api_key"

🔧 Tool List

1. WebSearch - Web Search

Functionality: Performs web searches to obtain chemistry-related information
Input: Search query string
Output: Search results and relevant links

2. MoleculeWeight - Molecular Weight Calculation

Functionality: Calculates the molecular weight of a molecule
Input: SMILES structure string
Output: Numerical value of the molecular weight

3. MoleculeAtomCount - Atom Count

Functionality: Counts the number of each type of atom in a molecule
Input: SMILES structure string
Output: Count statistics for each atom type

4. MoleculeSimilarity - Molecular Similarity

Functionality: Computes the Tanimoto similarity between two molecules
Input: Two SMILES structure strings
Output: Similarity score (0-1)

5. FunctionalGroups - Functional Group Identification

Functionality: Identifies functional groups present in a molecule
Input: SMILES structure string
Output: List of detected functional groups

6. SmilesCanonicalization - SMILES Canonicalization

Functionality: Converts a SMILES string to its canonical form
Input: SMILES structure string
Output: Canonicalized SMILES string

7. Iupac2Smiles - IUPAC Name to SMILES

Functionality: Converts an IUPAC chemical name to SMILES format
Input: IUPAC chemical name
Output: Corresponding SMILES string

8. Smiles2Iupac - SMILES to IUPAC Name

Functionality: Converts a SMILES format to an IUPAC chemical name
Input: SMILES structure string
Output: Corresponding IUPAC chemical name

9. Smiles2Formula - SMILES to Molecular Formula

Functionality: Converts a SMILES format to a molecular formula
Input: SMILES structure string
Output: Molecular formula (e.g., C2H6O)

10. Name2Smiles - Common Name to SMILES

Functionality: Converts a common chemical name to SMILES format
Input: Chemical name
Output: Corresponding SMILES string

11. Selfies2Smiles - SELFIES to SMILES

Functionality: Converts a SELFIES format to SMILES format
Input: SELFIES string
Output: Corresponding SMILES string

12. Smiles2Selfies - SMILES to SELFIES

Functionality: Converts a SMILES format to SELFIES format
Input: SMILES structure string
Output: Corresponding SELFIES string

13. Smiles2Cas - SMILES to CAS Number

Functionality: Converts a SMILES format to a CAS registry number
Input: SMILES structure string
Output: Corresponding CAS registry number

14. MoleculeSmilesCheck - SMILES Validation

Functionality: Validates the syntactic correctness of a SMILES string
Input: SMILES structure stringOutput: Validation Result (Valid/Invalid)

15. ReactionSmilesCheck - Reaction SMILES Validation

Function: Validates the syntactic correctness of a reaction SMILES string.
Input: Reaction SMILES string
Output: Validation result (Valid/Invalid)

16. MoleculeVisualizer - Molecule Visualization

Function: Generates a 2D structure diagram of a molecule.
Input: SMILES molecular structure string
Output: Molecular structure image

17. MoleculeCaptioner - Molecule Description Generation

Function: Uses AI to generate a textual description of a molecule.
Input: SMILES molecular structure string
Output: Detailed textual description of the molecule

18. MoleculeGenerator - Molecule Generation

Function: Uses AI to generate a molecular structure based on a description.
Input: Textual description of the molecule
Output: Corresponding SMILES molecular structure

19. MoleculeModifier - Molecule Modification

Function: Performs chemical modifications and derivatizations on molecules.
Input: SMILES molecular structure string
Output: Options for modified molecular structures

📖 Usage Examples

bash

Start MCP server with specific tools

python -m chemmcp --tools WebSearch MoleculeWeight MoleculeAtomCount

Or use all tools

python -m chemmcp --tools WebSearch MoleculeWeight MoleculeAtomCount MoleculeSimilarity FunctionalGroups SmilesCanonicalization Iupac2Smiles Smiles2Iupac Smiles2Formula Name2Smiles Selfies2Smiles Smiles2Selfies Smiles2Cas MoleculeSmilesCheck ReactionSmilesCheck MoleculeVisualizer MoleculeCaptioner MoleculeGenerator MoleculeModifier

🔧 MCP Configuration

Add the following configuration to your MCP client configuration file:

json
{
"mcpServers": {
"ChemMCP": {
"command": "/opt/miniconda3/envs/chemmcp/bin/python", // How to find the path: conda activate chemmcp && which python
"args": [
"-m", "chemmcp",
"--tools",
"WebSearch",
"MoleculeWeight",
"MoleculeAtomCount",
"MoleculeSimilarity",
"FunctionalGroups",
"SmilesCanonicalization",
"Iupac2Smiles",
"Smiles2Iupac",
"Smiles2Formula",
"Name2Smiles",
"Selfies2Smiles",
"Smiles2Selfies",
"Smiles2Cas",
"MoleculeSmilesCheck",
"ReactionSmilesCheck",
"MoleculeVisualizer",
"MoleculeCaptioner",
"MoleculeGenerator",
"MoleculeModifier"
],
"toolCallTimeoutMillis": 300000,
"env": {
"TAVILY_API_KEY": "your_tavily_api_key_here",
"LLM_MODEL_NAME": "openai/gpt-4o",
"OPENAI_API_KEY": "your_openai_api_key_here",
"CHEMSPACE_API_KEY": "your_chemspace_api_key_here", // Optional: Only used when additional database lookups are needed
"PATH": "/opt/miniconda3/envs/chemmcp/bin:/usr/local/bin:/usr/bin:/bin" // How to find the path: conda activate chemmcp && echo $PATH
}
}
}
}

Notes:

• How to Find the Correct Python Path:
  bash

  Activate your conda environment

  conda activate chemmcp

  View the Python path

  which python

  Copy the output path and replace the path in the above configuration

• Replace API keys with your actual keys.

• CHEMSPACE_API_KEY is optional: It is only used as a fallback database by the Iupac2Smiles tool if the PubChem lookup fails.

• Ensure that your Python environment version is 3.11 or higher.

🛠️ Dependencies

• RDKit: Core library for molecular processing and computation.
• PubChemPy: Interface to the PubChem database.
• SELFIES: SELF-referencing Embedded Strings for molecules.
• rdchiral: Chiral molecule processing.
• synspace: Molecule space exploration and modification.
• tavily-python: Web search API.
• litellm: Unified LLM API interface.
• requests: HTTP request library.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

Note: Some tools require corresponding API keys. Make sure to set the environment variables correctly before use.