Claude Code

Building AI Scientists with Claude Code and ToolUniverse

Overview

Claude Code integration enables powerful IDE- or terminal-based scientific research through the Model Context Protocol (MCP). This approach provides a developer-friendly interface for research while leveraging Claude’s advanced reasoning and ToolUniverse’s 600+ scientific tools.

┌─────────────────┐
│   Claude Code   │ ← IDE/CLI Interface & Reasoning
│                 │
└─────────┬───────┘
          │ MCP Protocol
          │
┌─────────▼───────┐
│ ToolUniverse     │ ← MCP Server
│   MCP Server     │
└─────────┬───────┘
          │
┌─────────▼───────┐
│ 600+ Scientific │ ← Scientific Tools Ecosystem
│     Tools       │
└─────────────────┘

Benefits of Claude Code Integration:

  • Developer Workflow: Use Claude inside VS Code/JetBrains or terminal

  • Advanced Reasoning: Claude’s strong multi-step reasoning

  • Comprehensive Tools: Access to 600+ ToolUniverse tools

  • Automated Execution: Natural-language to tools, directly in your editor

  • Batch & Iteration: Run multi-step research loops effectively

Prerequisites

Before setting up Claude Code integration, ensure you have:

  • Claude Code: Installed in your IDE or CLI

  • ToolUniverse: Installed

  • UV Package Manager: For running the MCP server

  • System Requirements: macOS, Windows, or Linux with Python 3.10+

  • API Keys: For specific tools or optional hooks (e.g., Azure OpenAI for summarization)

Installation and Setup

Step 0: Install Claude Code

Install Claude Code in your terminal (any OS with Node.js 18+):

# Standard (recommended)
npm install -g @anthropic-ai/claude-code

Verify and diagnose your installation:

claude doctor

Native installer (beta) alternatives:

# macOS/Linux/WSL: stable
curl -fsSL https://claude.ai/install.sh | bash
# latest
curl -fsSL https://claude.ai/install.sh | bash -s latest

On Windows PowerShell:

irm https://claude.ai/install.ps1 | iex

After installation, you can start Claude Code in a project:

cd your-project
claude

For details, see: Anthropic — Set up Claude Code.

For Windows installation, see: Anthropic — Windows setup.

Step 1: Install uv and ToolUniverse

Install uv package manager:

# macOS/Linux: official installer
curl -LsSf https://astral.sh/uv/install.sh | sh

# Verify installation
uv --version

For Windows installation and other methods, see: uv installation Tutorial.

Set up a dedicated uv environment and install ToolUniverse:

# Create a dedicated environment directory
mkdir -p /absolute/path/to/claude_toolu_env

# Install ToolUniverse into that uv environment
uv --directory /absolute/path/to/claude_toolu_env pip install tooluniverse

# Verify installation
uv --directory /absolute/path/to/claude_toolu_env run python -c "import tooluniverse; print('ToolUniverse installed successfully')"

Step 2: Test ToolUniverse MCP server

Test that the ToolUniverse MCP server works:

# Test the MCP server command
uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio --help

Step 3: Add ToolUniverse MCP server

Use Claude Code’s built-in command to add ToolUniverse as an MCP server:

# Add ToolUniverse MCP server with local scope (recommended for personal use)
claude mcp add tooluniverse --scope local --env AZURE_OPENAI_API_KEY=your-key --env AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio

Alternative scope options:

  • --scope local (default): Available only in current project directory

  • --scope project: Shared across project team via .claude/.mcp.json

  • --scope user: Available across all your projects

Environment variables (optional):

Add --env flags only if you need summarization hooks or tools requiring API keys:

# Minimal setup (no API keys needed for most tools)
claude mcp add tooluniverse --scope local -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio

Optimized Configuration for Research Users (Recommended):

# Add ToolUniverse with optimized settings for research
claude mcp add tooluniverse --scope local --env AZURE_OPENAI_API_KEY=your-key --env AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio --exclude-tool-types PackageTool --hook-type SummarizationHook

Configuration Benefits:

  • --exclude-tool-types PackageTool: Removes package management tools to save context window space if you don’t have coding needs

  • --hook-type SummarizationHook: Provides summary of the output that is too long to fit in the context window

  • AZURE_OPENAI_API_KEY and AZURE_OPENAI_ENDPOINT: Required for SummarizationHook functionality

Verify the server was added:

# List all MCP servers
claude mcp list

# Get details about ToolUniverse server
claude mcp get tooluniverse

See: Claude Code MCP documentation for advanced configuration options.

For MCP scope management, see: MCP installation scopes.

Step 4: Verify in IDE/CLI

After saving the configuration, verify connectivity:

  • Terminal (Claude Code CLI) - Launch in your project:

    cd /path/to/your-project
claude

    • In the chat, ask: “What ToolUniverse tools are available?”

    • If issues occur, run diagnostics:

      claude doctor

    • For terminal configuration, see: Claude Code CLI reference

  • VS Code - Restart VS Code, then open Command Palette and run: “Claude: Open Chat” - Ask: “What ToolUniverse tools are available?” - If tools don’t appear, check .claude/settings.local.json and reload window - For VS Code setup, see: Add Claude Code to your IDE

  • JetBrains (IntelliJ/PyCharm/etc.) - Restart IDE → open the Claude tool window - Ask: “What ToolUniverse tools are available?” - If tools don’t appear, review Tools → Claude Code → MCP Servers settings - For JetBrains setup, see: Add Claude Code to your IDE

Scientific Research Capabilities

Drug Discovery and Development

Claude Code with ToolUniverse enables comprehensive drug discovery workflows:

Target Identification: - Disease analysis and EFO ID lookup - Target discovery and validation - Literature-based target assessment

Drug Analysis: - Drug information retrieval from multiple databases - Safety profile analysis - Drug interaction checking - Clinical trial data access

Genomics and Molecular Biology

Access comprehensive genomics tools for molecular research:

Gene Analysis: - Gene information from UniProt - Protein structure analysis - Expression pattern analysis - Pathway involvement

Molecular Interactions: - Protein-protein interactions - Drug-target interactions - Pathway analysis - Functional annotation

Literature Research and Analysis

Comprehensive literature search and analysis capabilities:

Literature Search: - PubMed, Europe PMC, and Semantic Scholar - Citation analysis and trend detection

Content Analysis: - Abstract summarization - Key finding extraction - Gap identification

Multi-Step Research Workflows

Claude Code excels at complex, multi-step research workflows:

Hypothesis-Driven Research: 1. Formulate a hypothesis 2. Design an approach and select tools 3. Gather supporting evidence 4. Validate findings 5. Generate conclusions

Settings and Configuration

Tool Selection Strategies

Optimize tool usage for better performance:

Selective Tool Loading: - Load only relevant tools for specific research domains - Reduce context usage and improve response times

Example Tool Selection:

# Add ToolUniverse with specific tool filtering
claude mcp add tooluniverse-research --scope local -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio --include-tools EuropePMC_search_articles,ChEMBL_search_similar_molecules,openalex_literature_search,search_clinical_trials

# Verify the server configuration
claude mcp get tooluniverse-research

Multiple MCP Servers

Run multiple ToolUniverse instances for different purposes:

# Add research-focused instance
claude mcp add tooluniverse-research --scope local -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio --include-tools EuropePMC_search_articles,openalex_literature_search

# Add analysis-focused instance
claude mcp add tooluniverse-analysis --scope local -- uv --directory /absolute/path/to/claude_toolu_env run tooluniverse-smcp-stdio --include-tools ChEMBL_search_similar_molecules,search_clinical_trials

# List all configured servers
claude mcp list

# Remove a server if needed
claude mcp remove tooluniverse-research

Troubleshooting

Common Issues and Solutions

MCP Server Not Loading: - Verify ToolUniverse installation path and absolute paths - Check UV package manager installation - Run claude mcp list to see current servers - Check server logs with claude mcp get tooluniverse - For troubleshooting, see: Claude Code troubleshooting

No Tools Discovered: - Verify the ToolUniverse MCP server command runs locally - Check if your tool filters are too restrictive - Ensure all ToolUniverse dependencies are installed - Use claude doctor for system diagnostics

Tools Not Executing: - Provide required API keys via –env flags when adding the server - Verify network connectivity to external APIs - Check MCP output limits, see: MCP output limits

Tips

Tool Selection: Use –include-tools to load only the tools you need for better performance.

Status Check: Use claude mcp list and claude mcp get <server> to inspect MCP servers.

Keep Paths Absolute: Avoid relative paths in MCP config to prevent resolution issues.

Authentication: For OAuth-based MCP servers, use /mcp command in Claude Code chat for secure authentication.

Resources: Reference external resources with @server:protocol://path syntax in your prompts.

For comprehensive documentation, see: Claude Code documentation.