Model Context Protocol

Typed tool access
into real biology.

Every MCP server is a native extension of Claude's toolbelt — structured APIs into PubMed, GEO, GTEx, and more. Install once, and every chat session gets the tool. No API keys to juggle. No hallucinated accession numbers.

12+ curated MCPs One-click install OAuth-free
What's an MCP? Model Context Protocol is an open standard from Anthropic for giving models structured, typed access to external tools. Instead of Claude scraping HTML or faking API calls, each MCP exposes real endpoints with real return types. Operon bundles a curated catalog — and you can add any MCP server the community publishes.
Research MCPs

Twelve databases. One chat.

Each card is a real MCP server you can install from Operon's settings panel. Everything returns structured data Claude reasons over — not scraped pages.

PubMed

Literature

Full-text literature search grounded in NCBI's E-utilities. Cite real DOIs inline during any chat — no API key required.

search_papers fetch_abstract get_doi

BioMCP

Structural

PDB structure retrieval, UniProt lookups, AlphaFold access — everything you need for structural biology in one bundle.

fetch_pdb uniprot_lookup alphafold_get

GEO

Expression

Query any GSE / GSM accession, pull metadata, stream count matrices. Operon adds a caching layer so you're not re-downloading.

get_gse list_samples download_matrix

GTEx

Expression

Tissue-specific expression across the GTEx v8 release. Perfect for "is my gene expressed in this tissue?" questions.

gene_expression tissue_specificity eqtl_lookup

JASPAR

Motifs

Transcription-factor PFMs and PWMs for motif enrichment workflows. Plugs directly into your ATAC or ChIP pipeline.

get_motif scan_sequences enrichment

KEGG

Pathways

Pathway-level enrichment and gene-ontology crosswalks. Returns structured pathway maps you can drop into a notebook.

pathway_enrich get_pathway gene_to_pathway

AlphaFold

Structural

Predicted structures by UniProt ID, straight into PyMOL. pLDDT-aware confidence maps, aligned homologs, bulk download.

fetch_structure plddt align_homologs

Gene databases

Annotation

Cross-species ortholog lookup, Ensembl / RefSeq / HGNC resolution. Works with MyGene.info and biomaRt behind the scenes.

resolve_id ortholog gene_info

LatchBio

Workflow

Submit, track, and retrieve results from Latch's cloud-native bioinformatics platform — right from the chat.

submit_workflow get_status download_output

Filesystem

Core

Typed filesystem operations that respect .gitignore, sandbox to your project root, and work identically over SSH.

read_file write_file list_dir

Web fetch

Core

Controlled URL fetching with allowlists — pull a methods section, a GitHub README, or an open-access PDF without leaving the IDE.

fetch_url fetch_pdf extract_text

Notebook

Core

Read and edit Jupyter notebooks cell-by-cell. Claude can re-run a cell, inspect outputs, and patch a bug without re-running the whole thing.

read_cells edit_cell execute_cell
How it works

From install to first query in 90 seconds.

1

Open Settings → MCP

The settings panel has a dedicated MCP tab. Every bundled server is listed with a description and tool preview.

2

Toggle on, approve

Flip the switch. Operon auto-installs the server binary if needed and registers it with your Claude config.

3

Ask in chat

"Find recent papers on SPP1 in microglia." Claude picks the PubMed MCP, runs the query, cites the DOIs back.

Custom MCPs too. Paste any MCP server config (stdio, SSE, or HTTP) into the settings panel. Operon validates the schema and wires it up — same UI, same approval model.

Give Claude the right tools.

Install Operon — the MCP catalog is built-in. No extra setup.

Download Operon Read the docs