MCP Protocol

The Model Context Protocol (MCP) enables DB-GPT agents to connect with external tools and services through a standardized interface.

What is MCP?

MCP is an open protocol that provides a standard way for AI applications to connect with external data sources and tools. DB-GPT supports MCP as both a client (consuming MCP tools) and a server (exposing DB-GPT capabilities as MCP tools).

Architecture

Using MCP tools in agents

Step 1 — Configure MCP servers

MCP servers are configured in your TOML config file or through the Web UI's agent configuration.

Example TOML configuration:

[[agent.mcp_servers]]
name = "filesystem"
command = "npx"
args = ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/directory"]

[[agent.mcp_servers]]
name = "web-search"
command = "npx"
args = ["-y", "@modelcontextprotocol/server-brave-search"]
env = { BRAVE_API_KEY = "${env:BRAVE_API_KEY}" }

Step 2 — Assign tools to agents

In the Web UI:

1. Go to Apps → create or edit an app
2. In the agent configuration, select available MCP tools
3. The agent can now use these tools during conversations

Step 3 — Use in chat

When chatting with an MCP-enabled agent, the agent automatically selects and invokes the appropriate tools based on your request.

Supported MCP server types

Type	Description	Example
stdio	Local process communication	File system access, code execution
SSE	Server-Sent Events over HTTP	Remote APIs, cloud services

Common MCP servers

Server	Purpose	Package
Filesystem	Read/write local files	@modelcontextprotocol/server-filesystem
Brave Search	Web search	@modelcontextprotocol/server-brave-search
GitHub	Repository operations	@modelcontextprotocol/server-github
PostgreSQL	Database queries	@modelcontextprotocol/server-postgres
Slack	Send/read Slack messages	@modelcontextprotocol/server-slack

Finding MCP servers

Browse the growing ecosystem of MCP servers at the MCP Servers Directory.

DB-GPT as an MCP server

DB-GPT can also expose its capabilities as an MCP server, allowing other MCP-compatible applications to use DB-GPT features like:

• Knowledge base queries
• Database access (Text2SQL)
• Agent execution

Next steps

Topic	Link
Agent concepts	Agents
Agent development with tools	Tools Development
dbgpts community tools	dbgpts