# Predictive intelligence for drug discovery

The Predictive intelligence for drug discovery MCP server for Claude Code queries negative biomedical findings across seven modalities to check failed molecules and abandoned targets before lab testing.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · 1147 installs
- Author: nullary
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://smithery.ai/server/nullary/drug-discovery
- Updated: 2026-05-25
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/predictive-intelligence-for-drug-discovery

## What it does

This server connects an agent to failed experiment datasets across drug research. The index indexes 7 modalities, featuring 84.5M small molecules, 37.6M CRISPR screens, and 104K terminated clinical studies. Records come from ChEMBL, PubChem BioAssay, DepMap, ClinicalTrials.gov, and 10 other registries.

Every entry includes source provenance with DOI, PMID, and license details. Built-in tools like `search_inactive_compounds` and `search_target_history` report inactive candidate pairs, ADMET defects, or selectivity failures. You can also evaluate target fatigue with `get_target_landscape` using a gene identifier such as EGFR.

The integration relies entirely on a remote HTTP endpoint at `https://drug-discovery.run.tools`. No local datasets are required. Queries fail without an active internet connection to this host.

## Requirements
- Remote server: `https://drug-discovery.run.tools`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http drug-discovery https://drug-discovery.run.tools
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add predictive-intelligence-for-drug-discovery --url https://drug-discovery.run.tools
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "predictive-intelligence-for-drug-discovery": {
            "url": "https://drug-discovery.run.tools"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "predictive-intelligence-for-drug-discovery": {
            "type": "http",
            "url": "https://drug-discovery.run.tools"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http predictive-intelligence-for-drug-discovery https://drug-discovery.run.tools
```

## When to use it
- Look up inactive small-molecule compound-target pairs
- Check target attrition and past failures by gene symbol
- Track terminated clinical trials and CRISPR screen dropouts
- Review ADMET defects and drug interaction risks across modalities

## Example prompts
- Run `search_target_history` for EGFR to inspect failed attempts across all modalities
- Check inactive small-molecule candidates using `search_inactive_compounds`
- Inspect provenance and negative assay history for a compound with `get_compound`

## Similar
- [tooluniverse-gwas-drug-discovery](https://affpapa.org/en/nevibecode/skills/tooluniverse-gwas-drug-discovery.md) — skill · 392 installs
- [drug-discovery-search](https://affpapa.org/en/nevibecode/skills/drug-discovery-search.md) — skill · 169 installs
- [tooluniverse-small-molecule-discovery](https://affpapa.org/en/nevibecode/skills/tooluniverse-small-molecule-discovery.md) — skill · 240 installs
- [chembl-database](https://affpapa.org/en/nevibecode/skills/chembl-database.md) — skill · 2400 installs
- [chembl-search](https://affpapa.org/en/nevibecode/skills/chembl-search.md) — skill · 128 installs
- [Amass](https://affpapa.org/en/nevibecode/mcp/amass.md) — MCP server · ★ 1 repo stars

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
