Release Notes
Product Introduction
Purchase Guide
Getting Started
Buddy
- Buddy - Data Analytics Scenarios
- Data Engineering Agent
- Buddy Overview
- Data Governance Agent
- Buddy Data Security
- Overview of Buddy Agent Basic Capabilities
Platform Management
- Workspace Management
- Compute Resource Management
- Data Source Management Overview
Data Integration
- Supported Data Ingestion Solutions
- Supported Data Sources and Read Capabilities
- Data Source Management
- List of Data Sources
- Offline Data Synchronization
- Real-Time Data Synchronization
- File Upload
Data Engineering
- Studio
- Workflow
- CI/CD
Data Science
- Data Science Overview
- Model Experiments
- Feature Management
- Model Management
- Model Service
Data Analysis
- Data Exploration Overview
Data Governance
- Data Catalog Overview
- Data Discovery Overview
- Data Security Overview
- Data Quality Overview
- Semantic Model Overview
- Governance Tag Overview
Data Applications
- Agent Development
- MCP
- Model List
API Documentation
- History
- Introduction
- API Category
- Making API Requests
- Console APIs
- Workspace APIs
- Computational Resource APIs
- Development Studio APIs
- Workflow APIs
- Catalog APIs
- Data Types
- Error Codes
FAQs
Agreements
Contact Us
Glossary

What Is MCP?

Unduh

Mode fokus

Ukuran font

Terakhir diperbarui: 2026-10-09 16:17:13

Diterjemahkan oleh AI

MCP (Model Context Protocol) is an open protocol that exposes external tools, data sources, and business capabilities to large models / Agents through a standardized protocol. In DataBuddy, MCP serves as the unified bridge for Agents to call platform data assets. Agents can then discover and call these tools on demand in conversations or tasks, without writing separate adaptation code for each data source or service.
Here is a typical example: an intelligent data-query Agent needs to query the sales detail table, call the compute_gmv() function registered in Catalog, and then obtain real-time exchange rates through an external API. With MCP, you only need to register the sales detail Catalog functions and external HTTP APIs as MCP services, and the Agent can then call them uniformly as if they were local functions.
Features
Capability
Description
Unified protocol
Tools from different sources (Catalog functions / external HTTP / third-party plugins) are exposed to the Agent through the same protocol, so the Agent side does not need to understand underlying differences.
Multi-source access
It supports both platform built-in tools, user-defined tools, external MCP / API, and the MCP plugin marketplace, covering common enterprise integration scenarios.
Centralized Credential Management
Authentication for external MCP / API is centrally managed through Connection.
Seamless Integration with Agent
MCP services can be directly mounted to Agent debugging sessions, and call records are synchronized to MLflow Tracing.
MCP Service Types
DataBuddy Phase 1 supports the following MCP service sources, which you can centrally manage in the Agents > MCP list.
Category
Scenario
Creation Method
Catalog function (UC Function)
Call registered SQL / Python functions in Catalog (such as metric calculation and database queries) as tools.
After a function is created in SQL Editor or Notebook, it is automatically exposed as an MCP tool.
External MCP/API service
Connect to enterprise-owned MCP Server or third-party HTTP API.
In the MCP list, click Add External MCP Tool and configure it through mcp.json.
Partner MCP plugin marketplace
Install official MCP plugins with one click.
In the MCP list, click Discover More MCP Services from Marketplace to go to the plugin marketplace and install.
Key Concepts
To understand MCP, you need to be familiar with the following core concepts:
MCP Server: a set of externally exposed tools that corresponds to a callable MCP service address.
Tool: The smallest callable unit within an MCP service, containing a name, description, input parameter Schema, and output structure Schema. A single MCP service can contain up to 1000 tools.
Connection: Used to manage connection configurations and credentials for external MCP / API services.
MCP service address: a URL generated by the technical side that can be called by Agents / SDKs and supports copying.
MCP service status: indicates whether the current MCP is available. The list shows the status from the last time it was viewed and does not refresh in real time. The latest status check is triggered only when you enter the details page or click Refresh.
Tip
For a complete list of terms, see Core Concepts and Glossary.
Applicable Scenarios
Scenario
Applicable Method
Recommended MCP category
Agent queries business databases / calculates metrics
Register existing SQL / Python functions as MCP tools.
Catalog function
Agent calls existing internal APIs / MCP Server.
Custom MCP
External MCP/API service
Enable Agent to perform document extraction / web search / third-party data integration.
One-click installation with automatic parameter configuration
MCP plugin marketplace
Usage Process
A typical MCP usage loop consists of the following steps:
1. Create or install MCP services: Create Catalog functions and external MCPs in the MCP list, or install partner MCPs from the plugin marketplace.
2. View MCP details: Go to the details page to check the service status, copy the service address, and browse tools and parameter definitions.
3. Bind to Agent: Call MCP tools through the SDK in a high-code Agent.
4. Run and Observe: The entire process of Agent calling MCP is recorded in MLflow Tracing. You can view the input, output, duration, and status of each call in the Experiment module.
Sub-Feature Navigation
Sub-feature
Description
Documentation
MCP service configuration
Create, view, and manage various MCP services.
Configure MCP service.
MCP and Agent integration
Use MCP tools in Agent.
MCP and Agent integration
Recommended Learning Path
If you are using MCP for the first time, read the following sections in order:
1. Learn Core Concepts: this section.
2. Create your first MCP service: To configure an MCP service, the easiest way is to start with a Catalog function or by installing an ADP plugin.
3. Call tools in Agent: Integrate MCP with Agent.
4. Learn Agent capabilities: Agent Overview.
References
Agent Overview
Creating and Managing Agents
Configure Agent Tools and Knowledge Base
Model List Overview

Bantuan dan Dukungan

Apakah halaman ini membantu?

Anda juga dapat Menghubungi Penjualan atau Mengirimkan Tiket untuk meminta bantuan.

masukan

tencent cloud

Tencent Cloud DataBuddy

What Is MCP?

Features

MCP Service Types

Key Concepts

Applicable Scenarios

Usage Process

Sub-Feature Navigation

Recommended Learning Path

References

Bantuan dan Dukungan

Capability	Description
Unified protocol	Tools from different sources (Catalog functions / external HTTP / third-party plugins) are exposed to the Agent through the same protocol, so the Agent side does not need to understand underlying differences.
Multi-source access	It supports both platform built-in tools, user-defined tools, external MCP / API, and the MCP plugin marketplace, covering common enterprise integration scenarios.
Centralized Credential Management	Authentication for external MCP / API is centrally managed through Connection.
Seamless Integration with Agent	MCP services can be directly mounted to Agent debugging sessions, and call records are synchronized to MLflow Tracing.

Category	Scenario	Creation Method
Catalog function (UC Function)	Call registered SQL / Python functions in Catalog (such as metric calculation and database queries) as tools.	After a function is created in SQL Editor or Notebook, it is automatically exposed as an MCP tool.
External MCP/API service	Connect to enterprise-owned MCP Server or third-party HTTP API.	In the MCP list, click Add External MCP Tool and configure it through `mcp.json`.
Partner MCP plugin marketplace	Install official MCP plugins with one click.	In the MCP list, click Discover More MCP Services from Marketplace to go to the plugin marketplace and install.

Scenario	Applicable Method	Recommended MCP category
Agent queries business databases / calculates metrics	Register existing SQL / Python functions as MCP tools.	Catalog function
Agent calls existing internal APIs / MCP Server.	Custom MCP	External MCP/API service
Enable Agent to perform document extraction / web search / third-party data integration.	One-click installation with automatic parameter configuration	MCP plugin marketplace

Sub-feature	Description	Documentation
MCP service configuration	Create, view, and manage various MCP services.	Configure MCP service.
MCP and Agent integration	Use MCP tools in Agent.	MCP and Agent integration