Back to Browse

Biolab MCP Server

Developer ToolsLow Risk10.0MCP RegistryRemote
Free

Server data from the Official MCP Registry

Auditable MCP server for PubMed, Europe PMC, ClinicalTrials.gov, and bioRxiv/medRxiv queries

About

Auditable MCP server for PubMed, Europe PMC, ClinicalTrials.gov, and bioRxiv/medRxiv queries

Remote endpoints: streamable-http: https://srikarjy025-biolab-mcp.hf.space/mcp

Security Report

10.0
Low Risk10.0Low Risk

Valid MCP server (2 strong, 4 medium validity signals). No known CVEs in dependencies. Imported from the Official MCP Registry.

Endpoint verified · Open access · No issues found

Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.

Permissions Found in Source Code

Found by scanning the linked source code. This listing connects to a hosted endpoint, so none of this runs on your machine: it describes what the server software does where it is hosted.

file_system

Applies to the server that hosts this plugin, not to your machine.

HTTP Network Access

Connects to external APIs or services over the internet.

database

Applies to the server that hosts this plugin, not to your machine.

env_vars

Applies to the server that hosts this plugin, not to your machine.

How to Connect

Remote Plugin

No local installation needed. Your AI client connects to the remote endpoint directly.

Add this to your MCP configuration to connect:

{
  "mcpServers": {
    "io-github-srikarjy-biolab-mcp-server": {
      "url": "https://srikarjy025-biolab-mcp.hf.space/mcp"
    }
  }
}

Documentation

View on GitHub

From the project's GitHub README.

Biolab MCP Server

CI License: MIT Python 3.11+

Biolab is a small, local-first research workspace for finding papers, saving evidence, parsing selected papers, and asking questions over a paper-specific RAG index.

It keeps source responses, retrieval IDs, parser blocks, citations, and review decisions so results can be inspected later. It is a research tool, not medical advice or proof that a scientific claim is true.

Live demo: biolab-a2a.vercel.app

What you can do

  1. Search PubMed and other life-science sources.
  2. Create a project and select papers for it.
  3. Parse a selected paper into page-aware blocks.
  4. Review blocks and index accepted blocks in pgvector.
  5. Search the project RAG and save notes with citations.
  6. Download the parsed paper JSON.

The browser has two main pages:

  • / — dashboard for projects, search, and evidence.
  • /projects/<project_id> — one project’s papers, parser runs, RAG search, notes, and downloads.

Install

Requirements: Python 3.11+ and Git.

git clone https://github.com/srikarjy/biolab-mcp-server.git
cd biolab-mcp-server
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

Run locally

Start the dashboard and REST API:

biolab web

Open http://127.0.0.1:8000.

Start the MCP server instead:

python -m biolab.server

The MCP endpoint is http://127.0.0.1:8000/mcp.

Basic workflow

Search for papers:

biolab search "BRCA1 pancreatic cancer" --max 5

Or use the dashboard to create a project and select a retrieved paper. The project page then lets you parse it, review blocks, search the RAG, save notes, and download the parsed JSON.

Useful REST routes:

POST /v1/evidence/search
POST /v1/projects
GET  /v1/projects/{project_id}/report
POST /v1/projects/{project_id}/retrievals
POST /v1/projects/{project_id}/retrievals/{retrieval_id}/parse
GET  /v1/projects/{project_id}/parser-runs
POST /v1/projects/{project_id}/rag/search
GET  /v1/parser/runs/{parser_run_id}/download
POST /v1/projects/{project_id}/rag/notes

The download route returns an attachment such as paper.parsed.json. It contains parsed blocks and metadata. The original PDF is not retained by the parser endpoint.

Free local RAG

The free path runs embeddings on your computer with BGE-small. It does not need an OpenAI, DeepSeek, or paid embedding key.

Install the optional packages:

pip install -e ".[pgvector,local-embeddings]"

Start local pgvector:

export POSTGRES_PASSWORD=localtest
docker compose -f docker-compose.rag.yml up -d

Enable local embeddings:

export BIOLAB_LOCAL_EMBEDDINGS=1
export BIOLAB_LOCAL_EMBEDDING_MODEL=BAAI/bge-small-en-v1.5
export BIOLAB_EMBEDDING_DIMENSION=384
export PGVECTOR_DATABASE_URL=postgresql://biolab:localtest@127.0.0.1:54329/biolab

Run the vector integration test:

BIOLAB_PGVECTOR_INTEGRATION=1 \
pytest tests/test_rag_pgvector.py -m integration -v

Stop the local database when finished:

docker compose -f docker-compose.rag.yml down

Only blocks marked accepted are added to the project index. Search is scoped to the selected project and can be filtered by sections such as methods or results.

Tests

Run the normal test suite:

pytest -q

Run the RAG-focused tests:

pytest tests/test_rag.py tests/test_document_parser.py tests/test_inference.py tests/test_web.py -q

Run checks used by CI:

ruff check app.py biolab tests
mypy biolab

Storage and security

Local data is stored in biolab.db by default. Hosted and standalone APIs require authentication by default. Anonymous access is only for an explicit, isolated demo:

biolab web --allow-anonymous

The retrieval log is hash-chained for tamper evidence. It is not an immutable ledger and does not replace backups, access control, or scientific review.

For production deployments, configure a durable database and keep API keys in the deployment environment. Never commit credentials to this repository.

Reviews

No reviews yet

Be the first to review this server!