Log inSubmit server
Biolab MCP Server logo

Biolab MCP Server

Audits biological evidence searches across PubMed, Europe PMC, trials, protein, target, and preprint sources.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Biolab MCP Server

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "biolab-mcp-server": {
      "command": "uvx",
      "args": [
        "biolab-mcp"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

We couldn’t confirm it starts correctly

We ran the install command but it didn't respond within our test window — this can mean a slow first-time install rather than a real problem.

uvx biolab-mcp

No response to initialize.

This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked 11d ago.

Overview

Biolab MCP Server MCP server records and audits evidence retrieval from PubMed, Europe PMC, ClinicalTrials.gov, UniProt, Open Targets, and bioRxiv/medRxiv. It exposes MCP tools over hosted or local HTTP, stores source responses in SQLite or Turso, and returns retrieval IDs for later inspection. Reach for it when research workflows need traceable queries, preserved responses, projects, collections, watches, claims, or reports. It does not rank, interpret, or summarize evidence.

Use cases

  • Audit PubMed and biomedical literature searches
  • Preserve source responses for regulatory or research review
  • Search multiple biological evidence sources in one workflow
  • Track claims back to retrieval records
  • Monitor evidence with watches and collections

Key features

  • Audited retrievals with retrieval IDs
  • Hash-chained evidence records
  • Cross-source biological search
  • SQLite or Turso storage
  • Projects, collections, watches, claims, and reports

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 3d ago

Recent health check succeeded.

7/7 checks healthy over the last 46d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 10d ago via OSV.dev · biolab-mcp (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

Last commit14d agoMost recent push to the default branch.
Install checkInconclusiveDidn't respond in our test window — often a slow first install.
Directory activity4 viewsConfig copies, upvotes, and views on AllMCPs.

How Biolab MCP Server works

What Biolab MCP Server does

Biolab MCP Server MCP server provides an auditable layer between an AI agent and several biological information services. It sends searches to PubMed, Europe PMC, ClinicalTrials.gov, UniProt, Open Targets, and bioRxiv/medRxiv, preserves the returned source response, and gives the caller a retrieval ID. This makes it possible to inspect what was requested, when it was retrieved, and which source response supported a later claim.

The project is designed for evidence workflows rather than evidence interpretation. It does not rank results, summarize papers, or decide whether a scientific claim is correct. Instead, it records retrieval activity and supports organization around that record through projects, reviewed collections, evidence watches, claim links, and reports.

How it works

An MCP client calls a tool such as search_pubmed. The server then contacts the relevant external API, stores the response in a shared audit database, and returns the result with its retrieval identifier. Retrieval records use a hash-chained audit trail, making the stored sequence tamper-evident and inspectable later.

The local deployment uses SQLite. A production deployment can use Turso, a hosted SQLite-compatible database. The unified search_evidence operation can query PubMed, Europe PMC, ClinicalTrials.gov, UniProt, and Open Targets in one workflow while retaining explicit results and errors for each source. The standard search and fetch operations are also available in the 0.4 server described by the repository.

Setup and configuration

A hosted endpoint is available at https://srikarjy025-biolab-mcp.hf.space/mcp, so a client can connect without installing the project. The README gives configuration examples for Claude Desktop, Cursor, and Claude Code. Anonymous hosted use requires no signup, but shares a low-throughput budget; an authorization bearer key provides a separate, higher budget when available.

For local use, install Python 3.11 or newer and Git, clone the repository, create a virtual environment, and install the package. Running python -m biolab.server starts an HTTP MCP endpoint at http://localhost:8000/mcp. A local database file named biolab.db is created automatically. Docker is also supported; without Turso settings, container storage is ephemeral.

Tools and capabilities

The Biolab MCP Server MCP server includes source-specific search tools for PubMed, Europe PMC, ClinicalTrials.gov, and bioRxiv/medRxiv. PubMed retrieval stores full XML, Europe PMC also indexes bioRxiv and medRxiv content, ClinicalTrials.gov searches through its version 2 API, and bioRxiv/medRxiv searches support date-range pagination.

Other documented capabilities include:

  • Cross-source evidence search through search_evidence.
  • Standard search and fetch operations.
  • Retrieval inspection using stored retrieval IDs.
  • Projects and reviewed collections for organizing evidence.
  • Evidence watches for monitoring sources.
  • Claims and reports for connecting research records to conclusions.
  • UniProt protein records and sequences, plus Open Targets entities accessed through GraphQL.

Limitations and notes

External source availability and rate limits affect live searches. Anonymous hosted requests share a PubMed budget of approximately one request per second, while keyed access has a separate higher budget. The hosted deployment may not expose every capability in the repository's current local 0.4 version.

A local run avoids the hosted service's shared budget but still needs network access for live source queries. The default test command is designed to run offline; separate live tests contact PubMed, Europe PMC, ClinicalTrials.gov, and bioRxiv. Results are preserved for auditability, but the server itself does not validate scientific conclusions or replace expert review.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Biolab MCP Server

Biolab MCP Server is an MCP server that connects AI clients to PubMed, Europe PMC, ClinicalTrials.gov, UniProt, Open Targets, and bioRxiv/medRxiv. Its main tools search these sources, fetch or inspect retrievals, and organize evidence through projects, collections, watches, claims, and reports while preserving source responses for audit.

Install Python 3.11 or newer, clone the repository, create a virtual environment, and run `pip install -e ".[dev]"`. Start the local endpoint with `python -m biolab.server`; alternatively, install the published package with pip or pipx, or run the supplied Docker image.

Yes, connect an MCP client to `https://srikarjy025-biolab-mcp.hf.space/mcp`. The README provides setup examples for Claude Desktop, Cursor, and Claude Code. Anonymous access is available, but shared hosted rate limits apply.

Yes, Cursor can connect to the hosted HTTP endpoint or a local server URL. Add the endpoint under the `mcpServers` configuration as shown in the README, using the hosted URL or `http://localhost:8000/mcp` for a local deployment.

Local deployments use SQLite, with a `biolab.db` file created automatically. Production deployments can use Turso, a hosted SQLite-compatible database, by setting the documented Turso connection variables. Docker storage is ephemeral unless remote storage is configured.

Related MCP servers

All in Biology & Bioinformatics All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewBiolab MCP Server AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/biolab-mcp-server?style=directory)](https://allmcps.com/mcp/biolab-mcp-server)
HTML Embed
<a href="https://allmcps.com/mcp/biolab-mcp-server"><img src="https://allmcps.com/api/badge/biolab-mcp-server?style=directory" alt="Biolab MCP Server on AllMCPs" /></a>