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., "@T5Chem MCP Server Predict retrosynthesis for CCO."
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.
A Unified Deep Learning Model for Multi-task Reaction Predictions with MCP (Model Context Protocol) support.
Retrosynthesis Prediction: Predict reactants from a product molecule
Product Prediction: Predict products from reactants and reagents
Reagents Prediction: Predict required reagents for a reaction
Molecule Validation: Validate SMILES strings
Molecular Properties: Calculate detailed molecular properties
MCP Server: Integrate with AI assistants through Model Context Protocol
Related MCP server: MCP Server (fac_mcp_hub)
# Clone the repository
git clone https://github.com/bugatti742/t5chem.git
cd t5chem
# Install with MCP support
pip install -e ".[mcp]"
# Or install all dependencies
pip install -e .Large model files are NOT included in the repository. Download them separately:
# Download USPTO multi-task model
wget https://yzhang.hpc.nyu.edu/T5Chem/models/USPTO_MT_model.tar.bz2
tar -xjvf USPTO_MT_model.tar.bz2 -C model/Start the MCP server:
# Using default model path (model/)
t5chem-mcp
# Specify custom model path
t5chem-mcp --model_dir /path/to/your/modelpredict_retrosynthesis: Predict retrosynthesis routes
predict_product: Predict product from reactants
predict_reagents: Predict reagents for a reaction
validate_molecule: Validate SMILES strings
get_molecule_properties: Get molecular properties
# Batch prediction
t5chem predict --data_dir data/sample/reactants/ --model_dir model/
# Training
t5chem train --data_dir data/sample/reactants/ --output_dir model/ --task_type reactantsPython 3.10+
PyTorch 2.2+
Transformers 4.38+
RDKit 2022.9+
MCP SDK 1.0+
Jieyu Lu and Yingkai Zhang, J Chem Inf Model, 62, 1376 - 1387 (2022)
MIT License
Inputs: smiles (required).
Inputs: reactants_smiles (required), model_dir, num_beams, num_preds, +1 more.
Inputs: product_smiles (required), reactants_smiles (required), model_dir, num_beams, +1 more.
Inputs: product_smiles (required), model_dir, num_beams, num_preds.
Inputs: smiles (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 5 tools
Each tool targets a distinct chemical task: properties, product prediction, reagent prediction, retrosynthesis, and validation. There is no overlap in functionality.
All tools follow a consistent verb_noun pattern (e.g., get_molecule_properties, predict_product) with clear verbs and nouns, enhancing predictability.
Five tools cover the essential operations for a chemistry prediction server without being excessive or insufficient for the stated purpose.
Core workflows (property retrieval, forward/reverse prediction, validation) are covered. A tool for reaction conditions (e.g., temperature/solvent) is a minor gap but does not impede typical use.
Chemistry, pharma and biotech tools for AI assistants: NMR, MS, IR and UV analysis, ADMET and ICH M7, stability (ICH Q1), HPLC methods, DoE, docking, retrosynthesis, biologics. Deterministic results you can check. Built in Leipzig, Germany.
AI-powered bioprotocol optimization — generate, search, and manage lab protocols via MCP
Turn any LLM into your lab assistant: search samples, track experiments, analyze data with AI.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
