#MCP Servers

The Model Context Protocol (MCP) allows SciLaxy to connect to external tool providers. Add MCP servers to give your agents access to third-party APIs, databases, and services.

#What is MCP?

MCP is an open protocol for connecting LLM applications to external data sources and tools. An MCP server exposes a set of tools that agents can call, similar to builtin tools but hosted externally.

Examples of MCP servers:

• Database query tools (PostgreSQL, MongoDB)
• API integrations (GitHub, Slack, Jira)
• File system access
• Custom business logic

#Adding an MCP Server

1. Navigate to the MCP section in the sidebar
2. Click Add MCP Server
3. Configure the server connection:
   • Name — Display name for the server
   • URL — The MCP server endpoint
   • Transport — SSE or stdio
4. Save and test the connection

Via API:

curl -X POST /scilaxy/api/v1/mcps \
  -H "Authorization: Bearer {token}" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "GitHub Tools",
    "url": "https://mcp.example.com/github",
    "transport": "sse"
  }'

#Smithery Registry

SciLaxy integrates with the Smithery MCP registry, which provides a catalog of community-built MCP servers. You can browse and install MCP servers directly from Smithery:

1. In the MCP section, click Browse Smithery
2. Search for the tool or integration you need
3. Click Install to add it to your SciLaxy workspace

#Testing Tools

After adding an MCP server, you can test individual tools before using them in agents:

1. Open the MCP server details
2. Select a tool from the list
3. Provide test input parameters
4. Click Test to see the result

This helps verify that the server is properly connected and the tools work as expected.

#Using MCP Tools in Agents

Once an MCP server is added, its tools become available in agent configuration:

1. Open the agent settings
2. In the MCP Servers section, select the servers to enable
3. The agent can now call any tool from the selected MCP servers

MCP tools appear alongside builtin tools in the agent's tool list. The agent selects which tools to use based on the task at hand.