# chemistry [Health: Active]

**Category:** 🧬 Biology & Bioinformatics  
**Repository:** https://github.com/oliverkraft93-ops/covasyn-mcp-examples  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/chemistry

## Description
Deterministic chemistry MCP for AI agents — 130+ tools across 17 suites (ADMET, ICH M7/Q1).

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What chemistry MCP server does

The chemistry MCP server exposes CovaSyn's chemistry and analytics evaluations to AI agents through the Model Context Protocol. CovaSyn is aimed at pharma, biotech, and chemical laboratories, with evaluation areas that include impurity assessment under ICH M7, stability assessment under ICH Q1, spectra interpretation, and analytical method development.

The repository describes more than 130 tools organized into 17 suites. The tools are deterministic, and each result includes the method used and its limits. That information gives a subject-matter expert material to review and sign off rather than presenting an unexplained answer.

The repository also reports results from the MolecularIQ benchmark, which contains 3,540 symbolically verified chemistry tasks. The stated comparison is 14–41% correct without CovaSyn and 76–92% with CovaSyn. These figures are presented by the project as benchmark results, not as a guarantee for a particular workflow.

## How it works

CovaSyn acts as the underlying evaluation service, while this repository provides connection examples for MCP clients and custom applications. The examples cover Claude Desktop, Cursor, VS Code with an MCP extension, a minimal Python agent, a minimal TypeScript agent, and direct HTTP access with curl.

An agent sends supported chemistry requests through MCP and receives the corresponding CovaSyn evaluation. The returned method and limits can then be included in an agent workflow or checked by a domain expert. The repository is therefore primarily an integration and configuration reference; it does not describe a local chemistry engine included in the repository.

## Setup and configuration

Create an account through the CovaSyn workspace and generate an API key under the API keys section. Plans and terms are provided in that workspace. Then use the example appropriate to your client or application:

- Claude Desktop configuration
- Cursor MCP configuration
- VS Code settings with an MCP extension
- Python or TypeScript custom agents
- curl-based direct HTTP requests

The example snippets in the repository are MIT-licensed, while CovaSyn itself is commercial software operated and maintained by CovaSyn GmbH.

## Tools and capabilities

The chemistry MCP server is described as covering 17 tool suites and more than 130 tools. Supported areas named in the material include:

- ADMET evaluations
- ICH M7 impurity assessment
- ICH Q1 stability assessment
- Spectra interpretation
- Analytical method development

CovaSyn states that its compliance posture is designed for GxP environments, including EU Annex 11 and 21 CFR Part 11, with audit trail and EU data residency. Consult CovaSyn for the applicable service details before using it in a regulated process.

## Limitations and notes

Access requires a CovaSyn account and API key, and the service is commercial. The repository contains example connection snippets rather than a self-contained implementation of the underlying chemistry evaluations. It also does not document the complete tool catalog, request schemas, response schemas, rate limits, or API-key environment-variable conventions.

The listed compliance posture describes CovaSyn's stated design goals and service information; it should not be treated as an independent validation of a specific deployment. For managed evaluations or installation inside an organization's infrastructure, the README directs users to contact CovaSyn.

_Full upstream README: https://allmcps.com/mcp/chemistry/readme_