# Climb HLS MCP | Biomedical context for your Databricks agents
Source: https://climb.ai/labs/hls-mcp

The biomedical context layer for your Databricks agents

# Months of HLS data integration, now a one-click capability.

Climb HLS MCP brings ten curated biomedical and regulatory data sources to your Databricks agents on day one. Your agents get immediate, governed access to the public evidence your work depends on, composable with your own data, and nothing leaves your perimeter.

Install now Watch the demo

Governed by Unity Catalog Runs entirely in your perimeter

CLIMB.AI / HLS MCP Your perimeter

Public evidence

ChEMBL

Open Targets

PubMed

Trials

Agent

Your data

Lakehouse

Genie

Vector search

Unity Catalog

Reasoning over public + private data, nothing leaves the perimeter

10 data sources [ChEMBL](#src-chembl) [Open Targets](#src-open-targets) [PubMed](#src-pubmed) [bioRxiv / medRxiv](#src-biorxiv) [BioContextAI](#src-biocontext) [ClinicalTrials.gov](#src-clinicaltrials) [OpenFDA](#src-openfda) [CMS Coverage](#src-cms) [BioPortal](#src-bioportal) [US Census](#src-census)

Waypoints

[The Challenge](#challenge) [The Suite](#suite) [How It Works](#how) [Data Sources](#sources) [FAQ](#faq) [Get Started](#install)

[The Challenge](#challenge) [The Suite](#suite) [How It Works](#how) [Data Sources](#sources) [FAQ](#faq) [Get Started](#install)

The challenge

## The agents are ready. The data isn't.

Healthcare and life sciences leaders are racing to put agents on the decisions that matter, from molecular discovery and clinical trials to therapy coverage and safety surveillance. Yet very few have moved AI into production.

22%

Life sciences leaders who've scaled AI

9%

Achieving significant returns

Source: Deloitte, 2026 Life Sciences Outlook

The gap isn't ambition or model quality. The authoritative knowledge an agent needs is fragmented across a dozen public sources, each with its own integration burden, so teams spend quarters wiring up data instead of shipping outcomes.

The suite

## Biomedical context for every HLS workflow

We built Climb HLS MCP to eliminate that arduous integration work. Instead of wiring up a dozen public sources one at a time, you get a production-ready biomedical context layer for your Databricks agents, governed end to end and ready to reason over the public corpus alongside your own data.

### Discovery & translational R&D

Evaluate candidate molecules and their liabilities (ChEMBL), connect targets to diseases with supporting evidence (Open Targets), stay current with peer-reviewed and preprint literature (PubMed, bioRxiv), and ground answers in structured biology across genes, proteins, and pathways (BioContext.ai).

### Clinical development

Find and analyze active and completed trials for feasibility, competitive intelligence, and protocol design (ClinicalTrials.gov), and monitor approved-product safety, labels, and adverse events (OpenFDA).

### Commercial, market access & population health

Understand payer coverage decisions and the reimbursement landscape (CMS Coverage), standardize terminology across systems and data partners (BioPortal), and analyze population demographics for site selection, equity, and market sizing (US Census).

How it works

## How it works

Once installed, each server is registered as a Unity Catalog connection, governed by the controls you already run, and becomes available as tools across the AI ecosystem inside Databricks.

Your agents can access public biomedical evidence and enterprise data through a governed toolchain, resolved in a single call chain, inside your perimeter. The suite stays hosted and maintained, so there's nothing for you to operate.

49 tools Hardened auth MLflow traces Fully hosted

HLS MCP Architecture Open Interactive walkthrough Explore how the architecture works Expand to walk through the architecture

### For the platform team

-   Centrally discoverable and governed as Unity Catalog connections.
-   No bespoke integration and no shadow connectors.
-   No governance gaps.

### For research teams

-   Research cycles in weeks instead of quarters.
-   Decisions informed by the full evidence base.
-   Outcomes in this planning cycle, not the next.

Data sources

## The public corpus, built in

Curated biomedical, clinical, and regulatory sources, with 49 tools across one governed surface. Select any server to see what it brings to your agents.

Discovery & translational R&D

ChEMBL 2.4M+ bioactive compounds for drug discovery

Access 2.4M+ bioactive compounds and curated assay data from 90,000+ publications maintained by EMBL-EBI. Enables drug discovery teams to search compounds by name or structure, retrieve target bioactivity records filtered by IC50 or Ki, and build compound-target activity tables for hit identification and lead optimization.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/e370300d-57b5-4392-93df-50173f2c51ef/ChEMBL_ChEMBL-MCP-Bioactive-Compound-Drug-Discovery-Intelligence) [Data Source](https://www.ebi.ac.uk/chembl)

Open Targets Prioritize drug targets across genes, diseases

Access scored associations across 60,000+ gene targets and 20,000+ diseases aggregated from GWAS Catalog, ClinVar, ChEMBL, and FDA data. Enables pharma and biotech teams to systematically prioritize drug targets, identify clinical candidates, and build competitive pipeline landscapes in a single agent-driven workflow on Databricks.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/e1cf30e9-a64d-49d9-a21f-79cb7d3ef6a9/Open-Targets_Open-Targets-MCP-TargetDiseaseDrug-Intelligence) [Data Source](https://platform.opentargets.org/)

PubMed Search 35M+ biomedical citations and abstracts

Access 35M+ biomedical citations from MEDLINE and PubMed Central via NCBI E-utilities, updated daily. Enables life sciences teams to execute MeSH-guided systematic literature searches, retrieve structured abstracts and author metadata, and accelerate evidence synthesis, safety signal monitoring, and R&D landscaping workflows on Databricks.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/f36c2bde-856a-453c-bcfd-f6029d497519/PubMed_PubMed-MCP-Biomedical-Literature-Search-Evidence-Retrieval) [Data Source](https://pubmed.ncbi.nlm.nih.gov/)

bioRxiv / medRxiv Track life sciences preprints continuously

Access preprints from both Cold Spring Harbor Laboratory servers across life sciences and clinical medicine, updated continuously. Enables R&D, medical affairs, and scientific communications teams to detect emerging research signals ahead of journal publication, monitor competitor output, and support evidence synthesis and manuscript preparation.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/ab82287d-383f-4edf-906f-707c2127c728/bioRxiv-medRxiv_bioRxiv-medRxiv-MCP-Life-Science-Preprint-Intelligence) [Data Source](https://api.biorxiv.org/)

BioContextAI Knowledgebase Nine protein and pathway databases unified

Access nine specialized databases including AlphaFold, STRING, UniProt, Reactome, Human Protein Atlas, InterPro, PanglaoDB, PRIDE, and Europe PMC through a single federated connection. Enables computational biology and translational research teams to analyze protein structure, interaction networks, pathway biology, and proteomics data in unified agent workflows on Databricks.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/38a67489-ef4f-4052-91f2-e3760b473392/BioContext-AI_BioContextAI-MCP-Unified-Structural-Systems-Biology-Intelligence) [Data Source](https://biocontext.ai/)

Clinical development

ClinicalTrials.gov Search 500,000+ global clinical trials daily

Access 500,000+ global clinical trials from the U.S. National Library of Medicine's authoritative registry, updated daily. Enables clinical operations, medical affairs, and strategy teams to automate trial feasibility analysis, benchmark protocol design, evaluate patient eligibility, and monitor competitor pipeline activity in real time.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/11bbed6d-af19-4042-9acd-d8fb8a2b1f66/ClinicalTrialsgov_ClinicalTrialsgov-MCP-Clinical-Trial-Intelligence) [Data Source](https://clinicaltrials.gov/data-api/api)

OpenFDA Detect drug safety signals and recalls

Access 20M+ FAERS adverse event reports, FDA-approved drug labels, and recall enforcement actions from the open.fda.gov API, updated continuously. Enables pharmacovigilance, regulatory affairs, and drug safety teams to detect safety signals, review label content, and monitor recall activity directly within Databricks pipelines.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/23b9e8b4-4730-4cc5-8682-25881a3645cb/OpenFDA_OpenFDA-MCP-FDA-Drug-Safety-Labels-Enforcement-Intelligence) [Data Source](https://open.fda.gov/apis)

Commercial, market access & population health

CMS Coverage Verify Medicare coverage criteria and policy

Access National Coverage Determinations, Local Coverage Determinations, Medicare Coverage Documents, and Technology Assessments from the CMS Medicare Coverage Database. Enables payers, health systems, and revenue cycle teams to verify coverage criteria, monitor policy changes, compare jurisdictional variation, and support prior authorization and reimbursement analysis workflows.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/ee81cc46-64e7-4485-b3ef-f1584662df8a/Medicare-Coverage-Database_CMS-Coverage-MCP-Medicare-Coverage-Determinations-Policy-Intelligence) [Data Source](https://api.coverage.cms.gov/)

BioPortal (NCBO) Map clinical terms to standard vocabularies

Access 1,000+ biomedical ontologies including SNOMED CT, ICD-10, LOINC, and RxNorm from the National Center for Biomedical Ontology. Enables healthcare organizations to resolve free-text clinical terms to standard codes, map terminology across vocabularies, and standardize multi-source EHR datasets to controlled vocabularies at scale.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/b2a4c5c6-e4a9-439d-9a9d-a7aeea36e89c/BioPortal_BioPortal-MCP-Biomedical-Ontology-Terminology) [Data Source](https://bioportal.bioontology.org/api)

US Census ACS Population health and SDOH demographic data

Access U.S. Census Bureau population health and SDOH data at national, county, census tract, and zip code level from the American Community Survey. Enables health systems, payers, and public health teams to build SDOH risk profiles, map care access gaps, and support health equity reporting and CMS compliance.

[Databricks Marketplace Listing](https://marketplace.databricks.com/details/30fa8b86-366e-45dd-8db2-35e7ad03ce05/American-Community-Survey-ACS_US-Census-ACS-MCP-Population-Health-SDOH-Intelligence) [Data Source](https://api.census.gov/data.html)

FAQ

## Installation & setup questions

-   001 What permissions do I need to install a Climb MCP service?

    You must have permission to access and install products from the Databricks Marketplace within your Databricks workspace.

    If you do not see the Marketplace option or cannot install products, contact your Databricks administrator to request the necessary permissions.

-   002 How do I find a Climb MCP service in the Databricks Marketplace?

    1.  Open your Databricks workspace.
    2.  Navigate to Marketplace from the left navigation menu.
    3.  In the Marketplace search filters, select MCP as the product type.
    4.  Search for the desired Climb MCP service using one of the following names:
        -   `pubmed`
        -   `chembl`
        -   `census`
        -   `biocontext`
        -   `clinicaltrials`
        -   `bioportal`
        -   `opentargets`
        -   `biorxiv`
        -   `openfda`
        -   `cms-coverage`
    5.  Select the service you would like to install.

-   003 How do I install the MCP service?

    After opening the product page:

    1.  Click Install in the upper-right corner.
    2.  Enter a meaningful installation name.

        We recommend using a naming convention such as:

        -   `climb-pubmed-mcp`
        -   `climb-biocontext-mcp`
        -   `climb-openfda-mcp`

        This makes it easier to identify and manage MCP connections later.

    3.  When prompted for authentication, enter the Bearer Token that was provided in your onboarding email.
    4.  Complete the installation process.

    Once installation is complete, the MCP service will be available within your Databricks environment.

-   004 How do I verify that the connection was created successfully?

    After installation:

    1.  Open Catalog from the left navigation menu.
    2.  Click the gear icon (Catalog Settings).
    3.  Select Connections.
    4.  Locate the connection using the installation name you provided during setup (for example, `climb-pubmed-mcp`).

    You should see the connection listed and available for use by Databricks applications such as Genie, AI/BI Dashboards, notebooks, and agents.

-   005 Where do I find my Bearer Token?

    Your Bearer Token is included in the onboarding email sent by the Climb team.

    If you cannot locate the email or believe your token has expired, contact Climb support to request a new token.

-   006 Can I install multiple Climb MCP services?

    Yes. Each MCP service can be installed independently and provides access to a different external data source or knowledge corpus.

    For example, you may choose to install both `pubmed-mcp` and `clinicaltrials-mcp` to support biomedical research workflows.

-   007 I cannot see the Marketplace product. What should I do?

    Verify the following:

    -   You are logged into the correct Databricks workspace.
    -   You have Marketplace access permissions.
    -   The Marketplace filter is set to MCP.
    -   You are searching for the exact product name.
    -   The Bearer Token was entered exactly as provided.

    If the product still does not appear, contact your Databricks administrator or the Climb team for assistance.

-   008 Who should I contact for help?

    If you encounter installation issues, authentication errors, or have questions about MCP service configuration, contact the Climb team at [support@climb.ai](mailto:support@climb.ai).

Get started

## Empower your research with the full biomedical evidence base.

Your Databricks agents get governed access to the full biomedical, clinical, and regulatory corpus, reasoning over public evidence and your own data in the same workflow, all inside your perimeter.

Install now Watch the demo

 Sign up

## Get started with Climb HLS MCP.

A few details and we'll set up your access.

First name

Last name

Organization

Business email

 Keep me updated on Climb products, research, and events. You can unsubscribe anytime.

Sign up

## You're on the rope.

Thanks for signing up. We'll email your bearer token and connection details to the address you provided, and a Climb specialist will follow up shortly.

Done

Interactive walkthrough Close

Loading walkthrough…