Back to Documentation
AMADEV Docs
AmaStudio

AmaStudio MCP Guide

How MCP (Model Context Protocol) works in AmaStudio — built-in servers, configuration, and adding custom MCP servers.

AmaStudio MCP Guide

AmaStudio ships with several built-in MCP servers that give the AI model access to tools like filesystem, web search, database queries, and more.

Built-in MCP Servers

AmaStudio includes these built-in MCP servers:

  • Filesystem — Read, write, and manage local files
  • Web Search — Search the web via configured search providers
  • Database — Query connected databases
  • Shell — Execute shell commands in a sandbox
  • Memory — Store and retrieve conversation context
  • Knowledge Base — Query synced knowledge bases

Each server can be enabled/disabled from Settings → MCP Servers.

Adding a Built-in Server

  1. Open Settings → MCP Servers
  2. Click the + icon next to a built-in server in the list
  3. If the server requires configuration (marked with a yellow "Requires Config" tag), click the link to view setup docs
  4. Configure required fields and save

Custom MCP Servers

You can add your own MCP servers that follow the standard MCP protocol:

  1. Open Settings → MCP Servers
  2. Click Add MCP Server
  3. Choose the transport type:
    • stdio — Local process (command + args)
    • SSE — Remote server URL
  4. Enter the server configuration

stdio Example

Code
{
  "name": "my-custom-server",
  "type": "stdio",
  "command": "node",
  "args": ["path/to/server.js"],
  "env": {
    "MY_API_KEY": "your-key-here"
  }
}

SSE Example

Code
{
  "name": "remote-server",
  "type": "sse",
  "baseUrl": "https://example.com/mcp",
  "headers": {
    "Authorization": "Bearer your-token"
  }
}

MCP Tools in Conversations

Once configured, MCP tools are automatically available to the AI model during conversations. The model can call tools based on user requests — no additional setup needed.

The available tools depend on which MCP servers are active. Each server exposes its own set of tools that the model can invoke.

Troubleshooting

  • Server not appearing: Check that the server process starts correctly. Look for errors in the DevTools console (Ctrl+Shift+I)
  • Tool calls failing: Verify the server is running and any required environment variables are set
  • Timeout errors: Some tools may take longer — check the server's response time