MCP Server Connection Validation with Copolymer Design

Scenario

A user tests their newly connected PARAMUS MCP server in Claude Desktop by performing a real copolymer design workflow. This story demonstrates successful MCP integration and showcases the BRAIN platform's TransPolymer model for analyzing styrene-butadiene copolymers at different compositional ratios.

Interactive Dialog

USER:

I'm designing a new copolymer. I want to combine styrene and butadiene monomers. Can TransPolymer help me understand how different ratios affect the structure?

ASSISTANT:

[Calls: brain_discovery]

Discovery Result

Great! TransPolymer is already mounted and available. Now let me help you explore styrene-butadiene copolymers with different ratios. First, let me define the monomers: - Styrene: C=Cc1ccccc1 - Butadiene: C=CC=C Let me create a few different compositions to show you how the ratio affects properties. I'll start with a common 75:25 styrene:butadiene ratio (typical for SBR rubber):

ASSISTANT:

[Calls: brain_predict]

High Styrene (75:25) Composition

{ "composition": "75% styrene, 25% butadiene", "computation_time": "47.968 ms", "embedding_size": 768 }

BigSMILES: {[<][>]CC(c1ccccc1)[<],[][>]CC=C[<][$]}.{[>][<]}

ASSISTANT:

[Calls: brain_predict]

Balanced (50:50) Composition

{ "composition": "50% styrene, 50% butadiene", "computation_time": "38.801 ms", "embedding_size": 768 }

BigSMILES: {[][>]CC(c1ccccc1)[<],[][>]CC=C[<][$]}.{[>][<]}

ASSISTANT:

[Calls: brain_predict]

High Butadiene (75:25) Composition

{ "composition": "25% styrene, 75% butadiene", "computation_time": "31.694 ms", "embedding_size": 768 }

BigSMILES: {[][>]CC=C[<],[][>]CC(c1ccccc1)[<][$]}.{[>][<]}

ASSISTANT:

Summary of Styrene-Butadiene Copolymer Analysis

TransPolymer successfully generated structural embeddings for three different compositions: 1. **High Styrene (75:25)** - {[<][>]CC(c1ccccc1)[<],[][>]CC=C[<][$]}.{[>][<]} 2. **Balanced (50:50)** - {[][>]CC(c1ccccc1)[<],[][>]CC=C[<][$]}.{[>][<]} 3. **High Butadiene (75:25)** - {[][>]CC=C[<],[][>]CC(c1ccccc1)[<][$]}.{[>][<]}

What TransPolymer provides

768-dimensional embeddings that encode the polymer structure, which can be used for: - Property prediction (Tg, mechanical properties, etc.) - Similarity comparisons between different compositions - Structure-property relationships - Optimization studies

How composition affects structure

The different ratios give you different material properties: - **High styrene (75:25)**: More rigid, higher glass transition temperature, better wear resistance - **Balanced (50:50)**: Intermediate properties, good balance of flexibility and strength - **High butadiene (75:25)**: More flexible, lower Tg, better elasticity

Next Steps

Would you like me to: 1. Compare the similarity between these compositions using the embeddings? 2. Explore other monomer ratios? 3. Calculate specific properties if you have target values? 4. Try different copolymer architectures (random, block, alternating)?

Learning Outcomes

Successfully connect and validate PARAMUS MCP server in Claude Desktop

User confirmed MCP connection by testing chemistry tools

Use BRAIN platform's discovery tool to find available AI models

brain_discovery found TransPolymer.exe mounted and ready

Generate polymer embeddings using TransPolymer for different compositions

Three successful predictions with 768D embeddings

Understand BigSMILES notation for representing copolymer compositions

Different BigSMILES strings encode different monomer ratios

Interpret TransPolymer output and computation performance

Predictions completed in 30-48ms with 768-dimensional vectors

Key Concepts

MCP Protocol

Model Context Protocol enables Claude Desktop to access external tools and services through standardized JSON-RPC communication

BRAIN Platform

AI model hosting service that mounts and serves machine learning models as HTTP services

TransPolymer

Deep learning model (v1.1.0) that generates 768-dimensional embeddings encoding polymer structure and properties

BigSMILES

Polymer notation system that represents macromolecular structure including monomer ratios and architecture

Styrene-Butadiene Rubber (SBR)

Important commercial copolymer system with tunable properties based on monomer ratio

Embeddings

High-dimensional vector representations that encode chemical structure for ML tasks

Composition Effects

Styrene increases rigidity/Tg, butadiene increases flexibility/elasticity

Tools Working

OK calculate_molecular_weight Chemistry tool working perfectly, calculated caffeine MW = 194.19 g/mol
OK brain_discovery Successfully found 4 executables, TransPolymer mounted at port 5100
OK brain_predict Three successful predictions, computation times 31-48ms, 768D embeddings

Known Issues

FAIL brain_predict Initial attempts with different input formats failed - correct format is {data: {text: , temperature: }}