Ginkgo Datapoints, Tangible Scientific, and Inductive Bio Launch ADME-One™

   

Read More

Inductive Bio MCP Connector

Overview

Inductive Bio’s absorption, distribution, metabolism, excretion, and toxicity (ADMET) models predict the chemical and pharmacokinetic properties that determine whether a molecule can become a viable drug. By surfacing these predictions directly in your conversation with Claude, scientists can evaluate compounds, triage ideas, and prioritize which molecules to synthesize, without leaving their existing AI workflow.

Through this MCP server, users can access Inductive Bio’s models for the physicochemical properties LogD (lipophilicity) and pKa (acid/base ionization), key drivers of solubility and permeability. 

‍

Getting Started

  1. Go to Claude.ai > Settings > Connectors
  2. Find the Inductive Bio connector and click Connect
  3. Complete the OAuth sign-in when prompted in your browser
  4. Start a conversation with Claude - it can now use the Inductive Bio MCP tools

Alternatively, add the connector manually by running:

claude mcp add --transport http inductive https://api.inductive.bio/mcp

‍

Example Prompts

Drug discovery scientists use our MCP server to iterate on drug design ideas with Claude.

NSAID study: LogD (prompts appear in sequential order)

  1. Get LogD predictions for the top 10 reported NSAIDs. Which ones are most likely to be able to get into the brain?
  2. For ibuprofen, it has a low LogD. Decompose the molecule into fragments and retrieve predicted LogD for each part of the molecule - what part contributes most to its low LogD?
  3. Now that I know which part contributes the most to the low LogD, I designed some modifications to ibuprofen’s structure. Get predicted LogD values and tell me which modification is the best one to pursue if I want to get the drug into the brain. Here are the SMILES strings: 
    1. CC(C)CC1=CC=C(C(C(F)(F)F)C(O)=O)C=C1 
    2. CC(C)CC1=CC(F)=C(C(C)C(O)=O)C(F)=C1 
    3. CC(C)CC1=CC=C(C(F)(F)C(O)=O)C=C1 
    4. CC(C)CC1=CC=C(C(C)C(N)=O)C=C1 
    5. CC(C)CC1=CC=C(C(C)S(N)(=O)=O)C=C1 
    6. CC(C)CC1=CC=C(C(C)S(O)(=O)=O)C=C1 

ACE inhibitors study: pKa (prompts appear in sequential order)

  1. I'm interested in the predicted pKas of a few ACE inhibitors. Can you get the predicted acidic and basic pKas and tell me what that means for how well each drug is absorbed into the body?
    1. Captopril O=C([C@@H](CS)C)N1[C@H](C(O)=O)CCC1
    2. Enalapril O=C([C@@H](N[C@H](C(OCC)=O)CCC1=CC=CC=C1)C)N2[C@H](C(O)=O)CCC2
    3. Quinapril O=C([C@H](C)N[C@H](C(OCC)=O)CCC1=CC=CC=C1)N2[C@H](C(O)=O)CC3=CC=CC=C3C2
    4. Cilazapril O=C([C@@H](N[C@@H](CCC1=CC=CC=C1)C(OCC)=O)CCC2)N3N2CCC[C@H]3C(O)=O
    5. Ramipril [H][C@@]12CCC[C@]1([H])N(C([C@H](C)N[C@H](C(OCC)=O)CCC3=CC=CC=C3)=O)[C@H](C(O)=O)C2
    6. Lisinopril O=C([C@@H](N[C@H](C(O)=O)CCC1=CC=CC=C1)CCCCN)N2[C@H](C(O)=O)CCC2
  2. I designed some new molecules based on enalapril that I expect should improve absorption. Get the predicted pKa’s and LogD’s for these molecules - which one would you say to prioritize, compared to the predicted pKa and LogD for enalapril?
    1. O=C([C@@H](N[C@H](C(OCC)=O)CCC1=CC=CC=C1)C)N2[C@H](C(O)=O)C(F)(F)CC2
    2. O=C([C@@H](N[C@H](C(OCC)=O)[C@@H](COC)CC1=CC=CC=C1)C)N2[C@H](C(O)=O)CCC2
    3. O=C([C@@H](N[C@H](C(OCC)=O)CCC1=NC=CC=C1)C)N2[C@H](C(O)=O)CCC2
    4. O=C([C@@H](N(C)[C@H](C(OCC)=O)CCC1=CC=CC=C1)C)N2[C@H](C(O)=O)CCC2 

‍

Available Tools

list_available_models

List the molecular property prediction models you can run. Returns each model’s ID, assay name, and units. You can use the returned model IDs with the `predict_properties` tool.

predict_properties

Predict molecular properties for one or more molecules. Provide SMILES strings and the model IDs to run (from `list_available_models`); returns the predicted values per molecule and model.

‍

Privacy & Data

Inductive Bio does not retain any submitted chemical structures after requests complete. The MCP server does not collect any user prompts, inputs, or queries.

Inductive Bio does retain access logs only as long as necessary for security, compliance, and internal analytics. 

See the Inductive Bio MCP Access Privacy Policy  for full details.

‍

Contact & Learn More

Inductive’s full suite of models spans all tiers of ADMET assays, including microsomal stability, efflux, CYP inhibition, brain penetration, hERG inhibition, and more, and are fine-tuned to each customer’s chemical space to advance predictive performance. To access the full suite, reach out to Inductive Bio here or email mcp-user-support@inductive.bio.