Skip to content

Showcase: ApeRAG - Production GraphRAG Platform Built with mcp-agentย #426

@earayu

Description

@earayu

Hi mcp-agent team!

We built ApeRAG using mcp-agent as our core agent framework.

๐Ÿš€ Live demo: https://rag.apecloud.com/

Image Image

About ApeRAG

Production-ready RAG platform with Graph RAG, vector search, and intelligent AI agents.

How We Use mcp-agent

Agent Sessions:

from mcp_agent.agents.agent import Agent
from mcp_agent.workflows.llm.augmented_llm_openai import OpenAIAugmentedLLM

# Per-chat agent sessions with memory
self.agent = Agent(name="aperag_agent", server_names=["aperag"])
self.llm = await self.agent.attach_llm(OpenAIAugmentedLLM)

MCP App Factory:

from mcp_agent.app import MCPApp
from mcp_agent.config import Settings, MCPSettings

mcp_app = MCPApp(name="aperag_agent", settings=Settings(
    mcp=MCPSettings(servers={"aperag": MCPServerSettings(...)})
))

Event Processing:

from mcp_agent.logging.listeners import EventListener

class AgentEventListener(EventListener):
    async def handle_event(self, event: Event):
        # Route tool results to chat sessions by trace_id

Production Features

  • Multi-tenant sessions with isolated memory
  • Custom MCP server for knowledge base operations
  • Real-time tool processing with event routing
  • Internationalization (English/Chinese)
  • Enterprise error handling and retry logic

Scale

Currently serving thousands of users in production with mcp-agent powering our intelligent chat system.

Would love to be featured as an example! The project is open source (Apache 2.0).

Thanks for the amazing framework! ๐Ÿ™

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions