Skip to main content

MCP server for the Ersilia Model Hub

A Model Context Protocol server built on the official FastMCP SDK. It lets MCP clients search the Ersilia Model Hub catalog and fetch, serve, and run its AI/ML models, all over stdio.

Primitive Name Description
Tool search_model Searches the Ersilia model hub catalog by keyword.
Tool fetch_model Downloads a model to the local machine.
Tool check_model_fetched Reports whether a model has already been fetched.
Tool serve_model Starts a fetched model so it can accept predictions.
Tool generate_inputs Samples example inputs from a served model.
Tool predict Runs predictions with a served model and writes a CSV.
Tool close_model Stops a served model and frees its resources.
Tool delete_model Deletes a fetched model from local storage.

A typical workflow is search_model → fetch_model → serve_model → predict → close_model. You can optionally call generate_inputs after serving to sample example inputs to feed into predict, and delete_model to remove the model from local storage.

Installation

Ensure you have the following installed:

Clone the repository locally using git:

git clone git@github.com:ersilia-os/ersilia-mcp.git

Create a new virtual conda environment, activate it, and install the necessary packages:

conda create -n ersilia-mcp python=3.12
conda activate ersilia-mcp
pip install poetry
poetry config virtualenvs.create false --local
poetry env use $(which python)
poetry install

Client Setup & Registration

This MCP server has been tested mainly on Claude (specifically using Claude Code), but it supports any Model Context Protocol (MCP) host or model provider (such as Gemini or ChatGPT) that supports local stdio MCP servers.

Note on Claude Code: This repository is set up to be automatically configured with Claude Code out of the box. It includes a project-scoped .mcp.json configuration file so that when you run Claude Code in this workspace, the Ersilia MCP server is registered and started automatically.

Currently, we only have dedicated setup and registration documentation for:

We would love to receive documentation updates and setup guides for other MCP clients and platforms. If you have successfully set up this server with another client, please consider adding a new setup guide to docs/mcp_setup/! See CONTRIBUTING.md for contribution guidelines.

For further details on manual registration, debugging, or full development setup, please see DEVELOPMENT.md.

About the Ersilia Open Source Initiative

The Ersilia Open Source Initiative is a tech-nonprofit organization fueling sustainable research in the Global South. Ersilia's main asset is the Ersilia Model Hub, an open-source repository of AI/ML models for antimicrobial drug discovery.

Ersilia Logo

Metadata

Release files for ersilia-mcp 1.0.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for ersilia-mcp 1.0.1
File Size Uploaded
ersilia_mcp-1.0.1.tar.gz 21.6 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for ersilia-mcp 1.0.1
File Interpreter ABI Platform
ersilia_mcp-1.0.1-py3-none-any.whl Python 3 none any Details

Total release size: 48.9 kB

Release files / ersilia_mcp-1.0.1.tar.gz

Download URL ersilia_mcp-1.0.1.tar.gz
Size 21.6 kB
Tags Source
SHA-256 checksum
How to use checksums
099f97a15fabf126b9234935fe0054af9fb92bed737d670f664c3fffc6840184
BLAKE2b-256 checksum
How to use checksums
950a695113c7db98855e1ed41445be4e4415e6b135ce8fe8ff3e146bbadefb7e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via poetry/2.5.1 CPython/3.12.3 Linux/6.17.0-1022-azure

Release files / ersilia_mcp-1.0.1-py3-none-any.whl

Download URL ersilia_mcp-1.0.1-py3-none-any.whl
Size 27.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
20d59b65da5efda6acba565c59ba11d61609f82b94777a4ce9070ed55fd90ae7
BLAKE2b-256 checksum
How to use checksums
9371632badaff9486f810151f2ae01712e92f9862dcc89ecc70a177b0a48b7f0
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via poetry/2.5.1 CPython/3.12.3 Linux/6.17.0-1022-azure

Release history Release notifications | RSS feed

This release

1.0.1 This release

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page