Agent Framework

DB-GPT provides a data-driven multi-agent framework for building autonomous AI agents that can collaborate, use tools, access databases, and maintain memory across conversations.

Agent architecture

Every agent in DB-GPT is built around five core modules:

Module	Purpose
Profile	Defines the agent's role, name, goal, and constraints
Memory	Stores conversation history and learned information
Planning	Decomposes complex tasks into executable steps
Action	Executes tools, queries, and other operations
Resource	Provides access to tools, databases, knowledge bases

Key concepts

ConversableAgent

The base class for all agents. It implements the conversation loop: receive message, think (plan), act, respond.

Multi-agent collaboration

Multiple agents can work together on complex tasks:

• Sequential — Agents pass results to each other in order
• Parallel — Multiple agents work on sub-tasks simultaneously
• Manager-Worker — A planning agent delegates to specialist agents

Memory types

Memory Type	Scope	Persistence
Sensory	Current message	None
Short-term	Current conversation	Session
Long-term	Across conversations	Database
Hybrid	Combines all three	Mixed

Built-in agent types

DB-GPT includes several pre-built agents:

• Data Analysis Agent — Analyzes data, generates SQL, creates charts
• Summary Agent — Summarizes long documents and conversations
• Code Agent — Generates and executes code
• Chat Agent — General-purpose conversational agent

Quick example

from dbgpt.agent import ConversableAgent, AgentContext

# Define a simple custom agent
agent = ConversableAgent(
    name="DataAnalyst",
    role="You are a data analysis expert",
    goal="Help users analyze data and generate insights",
    llm_config={"model": "chatgpt_proxyllm"},
)

# Start a conversation
result = await agent.a_send("Analyze the sales trends for Q4 2024")

What's next

• Agent Introduction — Detailed agent framework guide
• Custom Agents — Build your own agents
• Agent Tools — Connect agents to tools
• Agent Planning — Task decomposition strategies